Woodlands

Item there is there one little park which contains in itself nine acres, with a laund, by the aforesaid perch. Which would be worth every year with the laund 4 shillings if there were no beasts. [Now the wood called Park Wood, 14.3 acres (56.8 ha). The laund would have been an open area, of which no trace survives.]

Item there is there one grove which is called tykele which contains in itself five acres by the aforesaid perch. [Now the wood called Titley Hill, 4.7 acres (1.9 ha).]

And a certain other grove which is called prestele which contains in itself thirty acres by the aforesaid perch. [Now Priestley Wood, 38.4 acres (15.5 ha).]

And another grove which is called Wetheresheg which contains in itself seven acres and a half by the aforesaid perch. And these three are worth every year sixteen shillings. [Now Swingens Wood, 14.2 acres (5.7 ha).]

Item there is there one great wood which is called boynhey which contains in itself ninescore acres by estimate. And it is worth per annum four pounds and ten shillings. [Now Bonny Wood, 126 acres (51 ha).]

[The length of the perch is given as 16.5 feet (5 metres), the same as the modern perch; the acres thus need no adjustment. Priestley and part of Bonny Wood are nature reserves of Suffolk Wildlife Trust.]

Conservationists for more than a hundred years were obsessed with preserving the world’s ‘primæval’ or ‘virgin’ forests. Until lately they meant forests unaffected by human activity (‘untrammelled by man’), or only by the sort of hunting, gathering and maybe rubber-tapping that seemed not to have much effect on the ecology.

Interest revived in the 1990s. Wildwood enthusiasts scolded conservationists for being concerned for coppices, wood-pastures and other ‘artificial’ kinds of vegetation. It was argued that even in Britain conservationists ought to turn all their energies to ‘restoring’ primæval vegetation, somehow unpicking the effects of thousands of years of management.

Ten years later, the old certainty, that wildwood meant forest, was challenged. If restorers of wildwood are to get anywhere they have to establish what it is they are trying to restore. The numbers, habits, tastes and effects of wild beasts are the most important unanswered question in prehistoric ecology.

The very concept of wildwood has shrunk in the face of archaeological and historical discoveries. Nineteenth-century explorers rediscovered the monuments of mighty civilisations in the jungles of Yucatán and Cambodia, grown up over the farmland that had supported these cultures. James Fairhead and Melissa Leach have demonstrated that much of the forest of West Africa was once savanna, maintained by a settled population. Even in the remotest reaches of the Amazon, clever people discovered how to make a living from the sodden, luxuriant, infertile vegetation: the present tranquil solitudes are not primæval but depopulated.¹

It can no longer be assumed that absence of evidence of ‘significant’ human presence means evidence of absence; that forests not containing temples and pyramids have always been forest. Faint earthworks, termite mounds, and traces of human cultures that did not build in stone are easily missed if one does not know what to look for. Archaeologists avoid forests because dense vegetation and leaf litter hide scatters of flints or potsherds. Anyone who restricts the term ‘natural woodland’ to woods with no human influence risks creating an empty category.

Old-growth forest is an American term for tracts that have escaped obvious logging or modern forestry practices within the life span of the existing trees (see Fig. 11). Some form the nuclei of National Parks; others come up from time to time for preservation or destruction, and are a political issue at which even Presidents shudder a little. These are wild and wonderful places, often containing very specific animals. It is easy to imagine them to be ‘virgin forest’ as it might have been had humanity never existed. But most were used by Native Americans and have been modified by the loss of their practices, as well as by the European habit of suppressing or postponing fires. George Peterken published a map of officially designated ‘virgin forests’ in Europe. Nearly half the sites were in the rather small country of Czechoslovakia (Table 10). Although this is partly due to a pioneering interest there in ‘virgin forest’ and its protection, it must mean that Czechs and Slovaks use a wider definition of ‘virgin forest’ than Poles, Germans or Hungarians. (Would Staverton Thicks in Suffolk or Ballochbuie Pinewood in Scotland count as ‘virgin forests’ in Czechoslovakia?)

According to Peterken, these tend to be designated according to whether they meet ecologists’, and especially foresters’, preconceptions of what a virgin forest should look like: ‘if they have large trees, much dead wood, a patchwork of structure, and a full range of dynamic states, the supposed state of [wholly] natural woodland’.² There has seldom been any serious attempt to verify that they were left unused by the surrounding human populations.

Peterken shows, and on my more limited experience I agree, that where a historical record exists it often demonstrates a history of wood-pasture. I once visited a gammelskog (‘ancient wood’) of huge, ancient spruces in Sweden; the local Mayor took me aside and pointed out some yet older pines of spreading shape which indicated that the spruces had infilled between what had once been scattered trees; Richard Bradshaw cited a pollen diagram showing that the site had once been a heath.

Sites left unmanaged for a century or more (of which England is beginning to acquire a goodly store as the years pass) are of the greatest ecological interest and should be protected. But they are not relicts of pre-Neolithic wildwood, and do not return to such with time. They are altered both by irreversible events in the more distant past and by events outside their borders. The very name of one in France, La Tillaie in the Forêt de Fontainebleau, proclaims that it was once a limewood, which could indeed have been a relict of Mesolithic times, but since then it has turned into an oakwood and then into a beechwood. A common sequence of events is a reduction of grazing as prehistoric people ate up the native herbivores; then a period of less tree cover and heavier browsing by domestic livestock in historic times; then a time of abnormally little browsing, after common-rights were suppressed; and lately resumed browsing as deer multiply.

Even such a seemingly feeble activity as fur-trapping can have an effect. In Ontario a major influence on forests is the food preferences of porcupines, which eat the bark off beeches and the leaves off hemlocks. If you ask why this happens in some areas but not others, you are told that it depends on the abundance of fishers, a kind of super-weasel, the only mammal that can digest a whole porcupine and live. The fisher has valuable fur on his back and in places has been kept down by trappers.³

The idea of wildwood, isolated from human activities, has so shrunk that one begins to ask whether it has existed at all since the last ice age, except on remote islands. The last extensive areas of wildwood could have been Madagascar and New Zealand, into which Polynesians burst about 1,500 and 800 years ago. On which lonely island did some storm-driven canoe or eighteenth-century ship bring the world’s very last wildwood to an end?

Neolithic culture in Britain is defined archaeologically by new styles of stone tools and the appearance of pottery. From pollen, cultivated cereals are distinguishable from other grasses. Another new feature is weeds, many of which came along with crops from southern Europe; plantains, however, were native. Small clearings can sometimes be identified, but a very distinctive feature of the beginning of the Neolithic is the Elm Decline. All over west and north Europe elm, and only elm, pollen suddenly and permanently declined.

A paradox of the forest model of wildwood is how and why Neolithic people should have abandoned hunting and gathering for the meagre yields of early cereals grown in competition with native vegetation. To do this at all they had to create fields and pastures. Why did they bring foreign crops and animals into an all-forest landscape that needed so much work before it became suitable for the new way of life? How did they find time and energy to dig up trees without metal crowbars and mattocks? What did they live on while doing it? On the Vera model the difficulty largely disappears: they would merely have to dig up grassland – leaving the wooded areas until farming was well established – and fence their cultivated plots to keep out tame and wild livestock.

In the seventeenth century Kentish clothiers and Buckinghamshire tailors abandoned good jobs to turn peasant in lonely, hostile, rattlesnake-haunted, poison-ivy-infested America. Somehow they found time and skill to make and fence fields as well as till them, but even the abundant records hardly explain what drove them or how they managed it. The problem is more severe for Neolithic farmers, unless they used land already cleared by wild animals – much as American colonists began on the savannas and farmland of their deceased Native American predecessors.⁴

The phrase ‘clearing forest by fire’ is a phantom that haunts discussions at this point. Here I dismiss it on the grounds that it cannot be done. To get an English wood to burn the trees have to be cut down, carried to a fire site, cut up, and densely stacked before a fire in one log will spread into the next. In the Norfolk mere deposits charcoal is even scarcer in the early Neolithic than at other times.⁵

Some of the remaining wildwood was converted to managed woodland, yielding trees small enough to be useful, whether for wood, leaves or timber. Evidence for coppicing is archaeological (p.220f); at this stage it is unlikely to be extensive enough to be discerned in the pollen record.

Elm – a moderate pollen producer – had amounted to about one-eighth of the trees of wildwood, but much more in some areas than others.⁶ At the beginning of the Neolithic it suddenly collapsed by more than half. This seems to have been simultaneous, rather than tracking the spread of Neolithic culture over Europe.

What was the cause? Deterioration of climate is inadequate to explain so universal a change. Had the climate become less favourable for elm, it would not have affected elm everywhere and elm only: it would have wiped out elm in climatically marginal areas, but not in the middle of its range. It should have affected elms in America too, but did not. The same applies to deterioration of soils.

An alternative, proposed in 1941 by Iversen in Denmark, is that people were feeding livestock on leaves, especially elm leaves, instead of grass, as they still do in parts of Europe. Elm would have been the favourite tree, and pollarding for leaves would have reduced its pollen output. The objection is that the small population of the time could not suddenly have taken to holding down the pollen production of such vast numbers of elms, even if they had nothing else to do. Nor was there a decline in alternative trees, such as ash, in regions where elm was scarce. The Elm Decline affected only elm, regardless of how common or rare it was.

The obvious explanation is Elm Disease, the effect of Ceratocystis fungi (p.336f). The Elm Disease epidemic of 1965 onwards was preceded in England by several others; there is no reason in principle why there should not have been one in 3800 BC. What other cause could have been so sudden, so specific to elms, so universal in Europe and yet absent from America (where Elm Disease was unknown until introduced in the 1920s)? At least one of the bark-beetle vectors is known to have been present.

The matter seems to be clinched by the pollen diagram from Diss Mere, where the sediment contains annual layers (varves) that allow the Elm Decline to be followed year by year. Here elms comprised about 10 per cent of the wildwood; they produced varying amounts of pollen in good and bad years; this ended in one very good year, after which all the years were bad. Ninety-four per cent of elm pollen production ceased in at most three to four years: exactly like a bad attack of Elm Disease.⁷

Elm was substituted by an increase of hazel, as in the 1970s when hazels were released from the shade of elms. Although pollens of different elms are almost indistinguishable, Keith Bennett was able to separate two types at Hockham and Stow Bedon Meres (Fig. 175), of which the type attributed to wych-elm predominates before the Elm Decline and that attributed to East Anglian elm after the Decline; this too would be in line with modern Elm Disease outbreaks, which affect some elms more than others. At Stow Bedon Mere there was little elm, and the Elm Decline was very subdued: this too parallels modern Elm Disease, where some patches of elm, especially in the depths of woodland, escape.

A symbiosis between Elm Disease and Neolithic people remains to be investigated. The behaviour of modern disease suggests that it would have been promoted by people felling trees and creating woodland edges, isolated trees and pollards. At the same time the gods, through Elm Disease, would have rewarded Neolithic people with fertile pastureland ready made.⁸

During the immense length of the Neolithic period, human settlement spread throughout the country. It was denser in some areas such as the Breckland, but seems to have been confined to river valleys in the claylands. Much of it was more than just clusters of huts in forest clearings. Most pollen diagrams show a great expansion of non-tree pollens. Monuments such as henges and long barrows involved precise alignments and called for a distant unobstructed horizon. Wide areas of what was later to be chalk downland and heath were already open country.

It used to be thought that ‘primitive’ people confined their activities to soils that were easily cleared and cultivated with their supposedly inefficient tools, especially the chalklands where Neolithic and Bronze Age monuments are so conspicuous. Forty years of archaeological survey and discovery have now shown how very pervasive prehistoric activity was in Britain, extending far into difficult terrain. The plateau of Dartmoor, now a place of punishment far above the altitudinal limit of modern settlement, is covered with Bronze Age houses and field systems. Most of the big wooded areas of medieval England contain traces of Iron Age and even earlier activity, for example the Neolithic chamber-tombs in Wychwood Forest. Prehistoric people could live where their successors, even in the medieval period of great pressure on land, did not.

Later prehistory gives the impression of a progressive advance of farmland, heath and moorland. Bronze and especially iron tools presumably made it possible to cut down trees faster than they could grow back up. Iron tools made it possible to till stony soils without having to get a new plough every day. Metal working created new requirements for fuel, stimulating the expansion of coppice-woods for conveniently small sizes of wood.

Lime declined more than other trees, but not everywhere simultaneously. At Diss Mere it disappeared in the mid- to late Bronze Age; in Epping Forest it went in the middle Anglo-Saxon period, although the site remained wooded. In some parts of the country lime is still undiminished in ancient woodland today. This whimsical decline (of a tree that is difficult to kill) is a mystery. It is all very well to say that prehistoric people harvested its leaves for fodder and its bark for fibre, or dug it up because it grew on the best soils, or that livestock ate lime in preference to beech, but if so, why did they attack it so unevenly?⁹

Woodless areas of the Bronze Age are marked by clusters of round barrows, placed where they could be seen from a distance, even if they were later to be in ancient woodland, as in the Forest of Dean or Wentwood Forest (Monmouthshire). Another indication of woodless areas is organised field systems. Field systems on the Dartmoor pattern – with bundles of parallel but sinuous axes in one direction subdivided by cross-walls at irregular intervals – are known in many parts of England and Ireland; some are in use to this day, including near Diss. They vary in date, where known, from Neolithic to Romano-British.¹⁰ Whatever the motive behind this kind of geometry, they indicate large areas of non-woodland. So do the vast field grids in Japan of the early centuries AD, such as the one within which the ancient capital city of Nara was later to be placed.

As woodland decreased and the landscape got more complex, the number of plant species increased: at Diss Mere only some 20 pollen types per sample are identifiable from wildwood times, increasing to 45 in the Roman period and 60 in the Middle Ages.

One begins to hear the claim that global warming has its roots far back in prehistory. Past concentrations of atmospheric carbon dioxide are known from fossil air trapped in polar ice cores. William F. Ruddiman notes that in the three previous glacial cycles CO2 rose during glaciations to about one and a half times its previous level, and then gradually fell during interglacials. In the Holocene the peak occurred early, at about 11,000 BC, but the fall, instead of continuing until the era of burning fossil fuel, ended about 6500 BC (early Neolithic in Asia) and was replaced by a steady rise. He attributes this to the spread of agriculture, to people grubbing out forests and indulging in practices, such as rice cultivation, that released carbon to the atmosphere that would otherwise have been locked up in vegetation and soils. He even ascribes subsequent falterings in the rise to plagues killing off humanity and allowing forests to recover.¹¹ In my view this is exaggerated: how could cultivation on what can hardly have been more than 6 per cent of the world’s land area (only 2 per cent of the whole globe) have been the cause of a 15 per cent anomaly in atmospheric CO2? However, other human activities, such as altering the frequency and extent of fire, could have contributed. If Australian Aborigines, by land management through burning, maintained savanna and prevented the return of rainforest over much of the continent, and if similar activities occurred in Africa and the Americas, the total area affected could have exceeded the area of cultivation.^fn1

The Roman Empire did not suddenly import a colonial culture unrelated to what had gone before, like the English in America; rather, it took over and added to indigenous activities, like the British in India. Expansion was especially in domestic heating and fuel-using industries: hypocausts, baths, bricks, tiles, glass and iron. Ancient Athens grew its food and fuel locally, and imported timber; Rome imported much of its food, and grew most of its timber and wood within Italy.¹² Roman Britain, as far as we know, did all these things within its own resources. The magnitude of some industries implies very extensive coppicing.

What was to be England, in Roman times, could hardly have been as much as 20 per cent woodland (as wooded as France today); some would say less than 10 per cent. Most of that woodland was managed. As yet there is no means of locating the woods or relating them to the structure of wildwood. Probably there was much more woodland in what is now Ancient Countryside than in future Planned Countryside. It is possible that some medieval woods already existed, but there is no way to identify them.

There have been wood-using industries since Neolithic times: some Somerset Levels trackways could have been built by trained carpenters. In the Middle Ages hurdles were usually made by specialists: there was a firm of hurdlemakers in the Hindolveston woods (Norfolk). Charcoal would have been made almost anywhere for local blacksmiths. As well as these dispersed, small-scale industries, there were heavy industries consuming enough wood to influence square miles of woodland.

The Romans had several such industries. Around Battle (Sussex) the iron industry appears to have produced 550 tons of iron a year in the period AD 120–240. If it took 12 tons of charcoal to make a ton of bar iron and 7 tons of wood to make a ton of charcoal, this would require 46,000 tons of wood a year, consuming the annual growth on about 23,000 acres (9,300 ha) of woodland. This calculation includes six out of the 36 Roman ironworking sites known in the Weald. The area of the Weald is 860,000 acres (348,000 ha). I guessed that in 1086 about 70 per cent of it was woodland, and at a further guess it was much the same in Roman times, giving 600,000 acres (242,800 ha) of woodland. The consumption figure may be an overestimate, for more recent experiments in making iron by Roman methods indicate a wood: iron ratio of 50:1 to 40:1.¹³ Wealden woods would have been amply able to support the known Roman iron-smelting; and even with further charcoal needed to process the iron there would have been land left over to feed the labour force and the horses and oxen.

There were smaller iron industries in the Forest of Dean and elsewhere. Another calculation for the whole of Roman Britain is based on a population of 3.6 million using 3.3 lb (1.5 kg) of iron per head, that is 5,400 tons per year. At a ratio of charcoal to finished iron of 20:1 and a wood: charcoal ratio of 7:1, this gives 800,000 tons of wood per year, roughly equal to the annual growth on 400,000 acres (161,880 ha) of woodland. I suspect this is an underestimate (especially as regards population). The total area of England and Wales is 37 million acres (15 million ha), of which somewhat under 2 per cent would have been devoted to growing fuel for the iron industry, plus further land needed to support about 150,000 workmen and their families.¹⁴ Although the total woodland area was probably much more than 2 per cent, it had to support several other industries as well as domestic uses. Roman woods are unlikely to have been over- or under-used.

I was taught, following the traditional view, that the Romans ‘made little impression upon the natural scene’. When their civilisation collapsed, most of the farmland was abandoned and reverted to woodland, from which the Anglo-Saxons slowly and laboriously carved out an independent cultural landscape. The ‘colonisation of medieval England’ continued even into the fifteenth century. If that were true, digging up trees should have been one of the best-known activities with which medieval men occupied their time. Familiarity with medieval documents showed that this was not so: there were records of assarts, farmland won from roughland, but most were of only an acre or two and some were from heathland or wood-pasture rather than woodland. The study of Domesday Book backdated the question to an earlier period: nine-tenths of the task had been accomplished before 1086.

Did Anglo-Saxons, then, spend two or three hours every day getting rid of trees? The entire corpus of Old English writing makes no direct mention of woodland being destroyed. No charter refers to land newly won from woodland; no boundary includes ‘the site of the wood that Æthelstan grubbed up’. The earliest records describe a cultural landscape already fully developed, with no suggestion that it was rapidly changing. That is not to say that woodland was never grubbed, but the inference is inescapable that Anglo-Saxons took over a previous landscape as a going concern.

This is supported by the pollen record and archaeology. Woodland certainly increased locally after the Roman period, for example on the site of Stansted Airport. But there is no sign of a general or overwhelming increase. The excellent record for Diss Mere shows a small increase of hazel and birch. At Quidenham Mere, only 7 miles (11 kilometres) away, but lying outside the organised field system around Diss, the main clearance phase seems to have been as late as the end of the Roman period.^fn2 An increase could occur well into the Anglo-Saxon period, as in Oxfordshire.¹⁵

Names of settlements give some indication of where early woodland was. I ignore ‘wood’ (as in Brentwood) because it has a very wide range of date and gives no indication of the amount of woodland. Useful names are those of settlements in -leah (‘wood’ or ‘clearing’, as in Rayleigh or Leigh-on-Sea), -hyrst (similar, as in Crowhurst), the Norse equivalent – þveit (as in Bassenthwaite), and -feld (‘open place in sight of woodland’, as in Bradfield). These are prevalent in areas where there was later much woodland, as in the Weald and south Lake District; they are rare in most of the future Planned Countryside. Kent – then as now – combined much woodland and a dense population, and is particularly rich in woodland place-names.¹⁶

The Anglo-Saxons, being fewer and far less industrialised than the Romans, would have used the woods less intensively, but did not allow them to spread far. Roman roads are an indication of the continuity of infrastructure, for they could not survive in disuse. Main roads, maybe, were kept open by the much reduced populations of the cities at either end. More significant are the Roman roads that survive in part: they ceased to be through routes, but there was still a local population that kept them up, year after year, as lanes between farms and hamlets.

Anglo-Saxon and early Welsh perambulations are the earliest documents to describe the landscape and to locate woodland. Already in the eighth century there were no vast and vaguely defined tracts of wildwood; there were wood-lots with owners and names of their own. Their boundaries were often marked by a linear feature called a wyrtwala ‘root-bank’ or wyrtruma. Something like one wood in four is still there, for example Ufton Wood (Warwickshire) (Fig. 45).

The division between the future Ancient and Planned Countryside is well marked in Anglo-Saxon charters, usually in features not directly related to woodland: hedges, open-field features, species of non-woodland trees. Woodland itself is more commonly mentioned in Ancient Countryside and often absent in future Planned Countryside, as in the Wiltshire chalklands and southeast Warwickshire.

The division between wooded and less wooded parts of England was established in or before the Roman period. In the more wooded parts the Anglo-Saxons took over Roman clearings and called them leys and hursts; later they may have made clearings of their own, but did not radically alter the distribution. In less wooded parts they probably conserved such woodland as there was. It was in those less wooded parts, with little room for expansion, that the Anglo-Saxons reorganised the cultural landscape by setting up open-field strips and by aggregating the earlier hamlets into villages.

What the Anglo-Saxons did with woodland is less well documented. The distinction between woodland and wood-pasture already existed. Some places possessed woodland at a distance, especially in the Weald where such detached wood-lots appear to have been mainly used for feeding pigs. As woodland specialists, they had at least ten words for different kinds of woodland. Charters mention various kinds of coppice products, charcoal, and woodland designated for fuel for salt-boiling. Woods were valuable property and could be the subject of lawsuits.¹⁷

William the Conqueror’s great survey of 1086 is a vast and mysterious record of land use throughout England, except the four northern counties and parts of the Welsh border. England then was not a very wooded land, even by twentieth-century standards. Only half the 12,580 recorded settlements possessed woods. Woodland was a common, but by no means essential, asset of an eleventh-century estate.¹⁸

For more than half the country Domesday gives the sizes of woods, and thus roughly their areas. For Eastern England woods are assessed in terms of the number of pigs they were supposed to feed, an unsatisfactory statistic that can be turned into areas only very approximately. For the southeast there is a still more nebulous assessment in swine-rents. Very seldom does Domesday name woods or say how many woods a place had.

The total area amounted to 15 per cent of the country. England thus had about twice as much tree cover (woodland, plantation, wood-pasture and non-woodland trees) as it has now. It was rather less wooded in Domesday Book than France is now. I first did the calculation 25 years ago;¹⁹ it has not been seriously challenged, though I suspect the figure of 15 per cent is a little too high.

Woodland was very unevenly distributed, as it is in France today (Fig. 176). The biggest concentration was in the Weald and the second biggest in the Chiltern plateau. Near Wolverhampton there was a tract of woodland some 15 miles (25 kilometres) across, though it had settlements within it. Larger areas of the east Midlands, Breckland, Fens, and east Yorkshire had no woodland at all. Elsewhere islands of woodland were scattered among farmland. Wood-lots ranged in size from 38 square miles (the future Cannock Chase, Staffordshire) to many of less than an acre.

The three distributions of woodland – from place-names, charters and Domesday – record its distribution early in the Anglo-Saxon period, later in the period, and just after its end. On the whole they are consistent, and point to the Anglo-Saxon as a period of general stability. They all independently make the distinction that later developed into Ancient versus Planned Countryside. Documents fail to mention people digging up trees because they did not often do it.

In 1086 the future Ancient Countryside had well over twice as much woodland in relation to its area as Planned Countryside. The two tracts differ. The southeastern tract, containing both the Weald and Chilterns, was one-quarter woodland; the northwestern tract only one-sixth. Much of the difference is due to the moorland in the northwestern tract. If moorland is excluded, Ancient Countryside had three times as much woodland in 1086 as Planned.

For Lincolnshire, Nottinghamshire and Derbyshire, Domesday separates coppice-woods from wood-pasture. Coppices tend to be smaller, and add up to 2 per cent of the total area of each county. This suggests that people encoppiced as much woodland as they needed and let the rest go to wood-pasture. Most of the ‘woodland’ area would thus have been wood-pasture, except in densely populated or sparsely wooded counties such as Cambridgeshire and west Suffolk.

Domesday says little about the composition of woods, though it occasionally mentions spinneys and carrs (woods of thorn or alder), and rarely oakwoods or willow-woods. It refers to industrial woods, such as those supplying the saltworks at Droitwich (Worcestershire) and Northamptonshire ironworks. It never mentions individual trees.

The earliest reliable county figures are for 1895 (Table 11). Most counties had more woodland in 1086 than 1895: Worcestershire had nine times as much. Leicestershire and Herefordshire, however, had not much more, and Devon apparently had less woodland in 1086. Generally regions with much woodland in 1086 lost more of it than regions that had little. Exceptions were the Weald and Chiltern plateau, which remain well wooded even now, although the woods are now less continuous. All England in 1086 had three times as much woodland as in 1895. The distinction between Ancient and Planned Countryside was still very marked in 1895, although the former had lost more woodland than the latter. The 1895 statistics, however, include plantations, by then not much less in area than natural woodland.

Comparing the 1086 figures with the area recorded as Ancient Woodland in the late twentieth century, between one-fifth and one-sixth of the Domesday woodland apparently survived until the 1930s, but only one-tenth was still intact (not grubbed or replanted) by 1990. (Most of the change was in the Locust Years 1950–75.) The distinction between Ancient and Planned Countryside is still visible; losses since 1950 have been greatest in Planned Countryside.

The survival rate is somewhat exaggerated, because figures for Ancient Woodland include woods originating between 1086 and c.1600; this, however, is offset by the inclusion as Intact of some woods that were successfully replanted before 1930. Losses since 1930 are overestimated, because much of the loss was to replanting, some of which is reversible (p.378ff).

After Domesday Book comes an active but poorly documented period, sparsely illuminated by charters and land-grants in monastic archives. When systematic documentation reappears in the mid-thirteenth century, much of the change to the modern distribution had already happened. Something like half the woodland of 1086 had turned into farmland or heath.

Under the rising population after 1086, farmland expanded to its sustainable limits and beyond. New land was won not only from the less-used woodland, but also from heath, mountain and the Fens (helped by low relative sea level). Most of the woodland on what is now Stansted Airport was assarted by the monks of Colchester Abbey and their neighbours. (For examples in Dorset.) Grazing animals might turn wood-pasture into heath by preventing the trees from replacing themselves, as in Thorpe Wood, whose owner, the Bishop of Norwich, could not prevent it from turning into Mousehold Heath in the twelfth century.²⁰ The few woods in southeast Warwickshire were more often conserved than the extensive woodland in northwest Warwickshire.²¹

Pressure on land also advanced the conservation and encoppicement of the remaining woodland. Woodland in formerly well-wooded areas became more intensively used and acquired scarcity value.

The change from wood-pasture or mere pannage to coppices sometimes came later in northern England.²² Growing industries of iron-smelting and ironworking maintained concentrations of coppice-woods. The growth of London as a market for fuel and timber came too late to save the great woods of Middlesex, but it did maintain an outer ring of woods in the northern Weald, Kent, Chiltern plateau and Essex. (Most of the woods of north and middle Surrey, however, are recent.) Paris, likewise, developed a system of supply and transport from distant woodland.²³

Norwich Cathedral Priory owned two woods at Hindolveston, north Norfolk. In 1272 the monks contrived to sell £214-worth of timber at once, a vast quantity. They needed the money because they had been sacked and burnt by the revolting townspeople (the fire-reddened stonework is still visible inside the Cathedral). In subsequent years they converted the woods to produce income from underwood rather than capital gains from timber. This involved especially attending to boundaries and security, and building woodbanks (p.159ff). This was a considerable capital expense, roughly one-quarter of the cost of a great barn.

By 1250 woods in England were fully developed. They had names, ownerships, boundaries and regular management, which were to remain for the next 700 years (Figs 177 & 179). Not that woodland was static: some woods were extended or curtailed, with new boundary earthworks recording the changes. Woods were grubbed out; new woods were formed as farmland or heath was abandoned. Even the many woods that did not change in extent had coppice cycles lengthened, rides cut through them, species added or subtracted, wild oaks replaced with planted oaks, and doubtless changed in other ways of which there is no record. Even if the underwood was more or less regularly cut, timber trees might be allowed to accumulate and then be suddenly felled to meet some disaster or to supply a big building project. However, all these are changes of detail, rather than affecting the continuity of the wood. Two big groups of woods illustrate some of the vicissitudes.

The great Abbey of Bury St Edmunds was endowed by Anglo-Saxon kings with about 221 estates: about half of West Suffolk, extensive tracts of northeast Suffolk and southeast Norfolk, and a few outliers. Of these estates 100 are known to have had woodland, recorded by Domesday Book in terms of swine. Of the others, 23 were in the woodless Breckland, and many others were very small. Where there was woodland in 1086, in 34 places it is not heard of again, and in two more it is not heard of after 1200 (Fig. 178).

The distribution of places that lost woodland is not random. In Norfolk and east Suffolk half the estates that had woodland in 1086 lost it by the later Middle Ages. The bigger woods disappeared: Mendham, with the biggest swine-entry of the Bury estates in Suffolk, is not heard of again. Chippenhall, the second biggest, is not heard of after the twelfth century; its ghost may, however, still be there as Chippenhall Green, now a splendid old grassland. In west Suffolk, however, only 11 out of 51 Domesday wooded places lost their woodland. A pattern emerges: the nearer a wood was to the Abbey the more likely it was to survive.

Maybe the monks had a deliberate policy of conserving the nearer woods and letting their tenants grub or graze away the more distant ones. The Abbey itself was a large user of wood and timber. Coppice-woods were becoming more valuable than arable land, and there were markets for timber and underwood in Bury and other west Suffolk towns, in the nearby woodless Breckland, and in the expanding economy of the woodless Fens. A hint of this is in a dozen places in west Suffolk where woodland is known to have existed, but is either not recorded in Domesday or gets a disproportionately small swine-entry. It is hardly credible that these woods came into existence between 1086 and 1250, especially as the Bradfield Woods, for example, contain no archaeological evidence of this. Probably they had already been encoppiced and were not used for swine; and since the form used by the Suffolk enumerators had no box for coppice-woods they were left out.

At the Dissolution in 1538 the Abbey owned about 150 separate woods. Some 17 woods (or the equivalent in fractions of woods) were destroyed between 1538 and 1700; about 30 in the eighteenth century; about 23 in the nineteenth (Fig. 180). Very little was lost between 1900 and 1950. Between 1950 and 1975 nearly one-third of the remaining woods (by number) were grubbed or replanted. About 38 appear to survive today either intact or coniferised, with fragments of a dozen others. Compared to other Abbey possessions, the woods have survived relatively well, but big woods have fared worse, especially in the twentieth century. In 1538 the Abbey had 13 woods of more than 100 acres; by 1950 six were still reasonably intact; today the only intact survivor is Felshamhall Wood (Bradfield Woods), nearly the smallest of the 13.

The endowment of St Etheldreda’s Abbey at Ely was on a similar scale to Bury, with about 225 estates, comprising most of the Isle of Ely, much of the Breckland and Norfolk Fens, and outliers extending into Hertfordshire and Essex. Only 52 of these estates had woodland in 1086. When the Pope founded Ely diocese in 1109, the endowment was split between the Bishop and the Abbey. The Bishop got 58 estates, but these included most of the bigger ones and those with woodland.

The Ely Coucher Book of 1251 records existing woods, and former woodland as assarts and as fields called Stocking (a place of ‘stocks’ or tree-stumps). On the estates round Somersham (Huntingdonshire), where the Bishop had a Forest, about 900 acres (364 ha) of woodland remained in 1251; stockings and assarts add up to 389 acres (157 ha) of arable land, whereas the Bishop’s arable land not remembered as former woodland came to 530 acres (214 ha). The general impression is that wooded estates lost about one-third of their woodland between c.1000 and 1251, whereas woodless estates (with rare exceptions) had no memory of former woodland. The pattern of survival is concentrated on Cambridge, not Ely.

After 1251 the rate of destruction slowed. The survey ordered by the king in 1356, when the Bishop was wanted for murder, generally records much the same woodland as in 1251.

Of 49 woods in the Ely Coucher Book, about 17 still survive as a whole or in substantial part, plus fragments, especially boundaries, of others. Big woods and especially common-woods disappeared: only two of the survivors, Hayley Wood and Bonny Wood (Barking, Suffolk), are bigger than 100 acres (40 ha). Six Coucher Book woods are nature reserves or belong to the Woodland Trust.

This pattern of slow attrition of woodland, but only minor changes in the woods that survived, is probably typical of areas with relatively little woodland and only ‘domestic’ uses of woods. Heavy industries bring booms and busts into the picture: the rise and collapse of an industry both have to be considered.

In the Middle Ages some industries operated at a similar scale to those of Roman Britain. Baths and hypocausts were forgotten; bricks and glass were being rediscovered; but the population was on a similar scale, and presumably dropped pots and wore away ploughshares at a similar rate. Ironworking was mainly in three of the biggest concentrations of woodland: the Weald, Forest of Dean and southern Lake District.

It used to be claimed that industries caused ‘deforestation’ – converting forest to non-forest. This rests on three misapprehensions: that fuel-using industries burnt timber; that they made no provision for future fuel supplies; and that trees once felled never grew again. Timber can, at a pinch, be used as a fuel or made into charcoal, but it would not chark all the way through unless laboriously chopped into billets. (The wages of charcoal-burners often came to more than the cost of trees or of felling them.) Here, as ever, the shoots from a stump were more useful than the original tree. Then as now, an industry was an investment in buildings, furnaces, equipment, manpower, horses etc., not to be jeopardised by letting the fuel supply run out. Industries tended to settle in large areas of woodland because charcoal is fragile and more difficult to transport than other raw materials.

calls for only one-sixth of a ton of carbon to make a ton of iron; the rest of the charcoal is spent in creating the high temperature needed, and can be saved by more economical furnace design. By 1550 England was said to produce about 5,000 tons of iron a year, much as in Roman times, but the wood: iron ratio was reduced to 30:1. This would be roughly equivalent to the annual wood yield from 75,000 acres (300 square kilometres) of woodland (on the basis that no timber was being grown). By 1700 production increased to 24,000 tons a year, but improved blast furnaces used wood at about 18:1, so the woodland needed was 220,000 acres (900 square kilometres), about one-seventh of all the woodland in England.²⁶

Industrial and domestic uses clashed, since industrialists, with many bigger bills to pay than that for trees, could buy up woods at prices that domestic users could not afford. By the fifteenth century, most of the woods around Sheffield were producing charcoal, much of it for making blades and nails. Even into the nineteenth century charcoal was used for making the best steel.²⁷

London depended on a great ring of woods outside its agricultural and market-gardening belt. On the south side, approaching the Weald, the conflict with the expanding iron industry became an affair of state, in which a statute of 1580 tried to prevent ironworks from using charcoal made within 22 miles (35 kilometres) of outer London.

Another clash was with the Cornish tin industry. Tin was more economical of fuel than iron, and less particular about which fuel, although charcoal was preferred. A ton of tin called for roughly ten tons of wood, and at its peak, c.1700, the Cornish tin industry would have been using about 15,000 tons of wood a year or its equivalent. Although this should have been within the capacity of the 30,000 acres (12,000 ha) of wood in Cornwall, these woods were already fully used, and as the industry moved westward from Dartmoor to Land’s End it bought up the meagre local woods. Around the Helford River:

… their few parcels [of woodland] yet preserved are principally employed to coaling, for blowing of tin. This lack [of domestic fuel] they supply either by stone coal fetched out of Wales, or by dried turfs.

The port of Gweek, in the midst of the Helford River woods, was importing charcoal.²⁸

Industrialists expanded into Wales, Scotland and even Ireland, not because they had destroyed the woods of England, but because they were expanding production. In Wales a sixteenth-century bard denounced the felling of the great wood Coed Glyn Cynon in the area later to be Merthyr Tydfil and Aberdare, where an iron industry was expanding.²⁹

The owners and occupiers of land are often averse to give correct statistics on a matter which might afterwards affect their interests.

Scotland is another country, and even the nature of woodland is not the same as in England. Woods less consistently have sharp edges. Pinewoods and birchwoods have a history of moving about within the same general area, but even the oakwoods were often not sharply demarcated from moorland.

In the post-medieval period, coppicing seems not to have been universal, developing chiefly through the influence of industries. Banks and walls around woods were apparently made from the seventeenth to the nineteenth century where there had previously been no fixed boundary.

At the same time, many woods were used as pasture and shelter for livestock. This must mean that the trees were not continuous but widely spaced enough to allow a worthwhile growth of grasses and other edible plants between them: they would thus have counted as wood-pasture rather than woodland.

This emerges from the researches of Professor Smout and his colleagues for the period after 1590.³⁰ Whether it is also true of the Middle Ages is less clear. As in England, many medieval Scots consumed timber and underwood, and (especially in the Lowlands) would have had cause to conserve the sources. However, Scotland had more alternatives, especially coal and peat. Even timber did not have to come from trees; it could be imported from Scandinavia, or dug up as fossil ‘bog fir’ from peat, or picked up on Hebridean beaches having drifted from an unknown western continent.

Industrialisation spread to Scotland. Greedy or impoverished lairds began leasing woods to optimistic English and Irish capitalists for various industries. Ironworks were among the most successful, beginning at remote Loch Maree in 1611. One of the latest and longest lived, the huge Lorn Furnace at Bonawe in Argyll, opened in 1752 and drew its supplies from oakwoods within a 50-mile (80-kilometre) radius, mostly by sea (which made transporting charcoal easier).

Decline was not, as many writers have assumed, due to English or Irish tree-fellers. On the contrary, the woods that were exploited are, for the most part, those that are still there: the disappearances were among those woods not known to have been exploited. In the post-medieval period conservation practices, such as keeping livestock out of woods after felling, were practised only sporadically unless there was a prospect that the wood might have more than a local commercial value.

In 1750, according to Smout, woodland occupied nearly 10 per cent of Scotland, falling to 3 per cent by 1900 (excluding plantations). I suspect that the former figure stretches to the limit everything that could come within the elastic Scottish definition of woodland, including scattered trees and bitten-down bushes. On this basis, the survival rate of Scottish woodland was less than in England, but more than in Ireland.

I agree with Smout in attributing the disappearance of Scottish woodland to these causes:

Continuing natural succession from woodland to peatland in the Holocene.
Increased grazing pressure in the nineteenth century: the substitution of multitudes of sheep for modest numbers of cattle and goats.
Human depopulation, reducing the motive to conserve.
Short-cycle burning on grouse moors, restricting the scope of birch and pine to move around.
Encroachment by modern forestry, turning woods into plantations.
Twentieth-century proliferation of deer.

None of these processes was uniform. It would be worth studying what happened to woodland in glens that escaped Highland Clearances, or on estates that promoted grouse versus deer.

The Old Irish texts, according to Professor F.S. Kelly, nowhere suggest that there were large areas of unused woodland in medieval Ireland. The impression is of a wood-based economy, using many smallish woods. The few medieval surveys and cartularies do not suggest much woodland.

The Irish Civil Survey of 1654 gives areas of thousands of woods, located by townland (subdivision of a parish) but seldom named. The published volumes of the Survey cover just over half of Ireland, in which the woodland adds up to 2.7 per cent of the land area (Table 12). In an earlier version of this calculation I added in seven more counties with no published volumes, but for which Eileen McCracken, Irish woodland historian, cited woodland areas from Survey manuscripts. All these counties had 1 per cent of woodland or less, making the recorded total for Ireland 2.1 per cent of the land area.³¹

I now think that McCracken’s figures are much too low (even though she maintained that Ireland had much more woodland than this). Her counties include Offaly and Leix, for which a famous map of 1562, although small-scale and vague, shows them to have been well wooded;³² and Kerry, which now has the biggest concentration of ancient woodland in Ireland. For nearly one-third of Ireland the Civil Survey seems not to survive at all; this includes what were probably the well-wooded counties of Wicklow and Fermanagh.

I conclude that the Civil Survey information is somewhat biased towards poorly wooded counties by the survival of the data. Allowing for this, the true total in the 1650s is likely to have been somewhat higher, perhaps 3.5 per cent. Ireland would still have been less wooded than England at the time.

What happened to the Irish woods? By c.1845, when the Ordnance Survey mapped every wood in Ireland, the woodland area was down to 1.5 per cent (albeit probably with a restrictive definition of woodland: see the Eagle’s Rock story). However, most of these woods were in different places. It is remarkably difficult to identify the same wood in the Civil Survey and the Ordnance Survey, less than 200 years apart. (Would it be easier if the Wicklow and Kerry data had survived?)

Sir Walter Raleigh, who had lifted some lands with the sword from the great Desmond Estate, commissioned a map of Mogeely, on the Cork–Waterford border, in 1598: an exceptional map, drawn to English standards of cartography, with the boundaries of every wood (Fig. 34). It was an agricultural countryside, with no visible bogs or mountain, big enclosed and named fields, and the remains of a big deer-park. About one-fifth of the land shown is woodland. Unknown to the cartographer, it was a land of ancient agriculture: there are several raths (Iron Age farmsteads) in the area mapped.

By 1844 there was a remarkable transformation. Nearly all the woods had vanished; so had most of the infrastructure. It had become a landscape of small fields and a few small woods (with un-Irish names like Belvidere Wood). There is little correspondence in the location, and none in the boundaries, of woods between 1598 and 1844. Only the main road and the village of Curraglass remained to identify the spot, together with the river Bride (which had lost its islands) and the church (in ruins).

One should not make too much of one map, although it is said to be the only surviving map of its type. If it is representative, then Irish woods were victims of a huge reorganisation of landscape, probably in the eighteenth century, without parallel either in England at the time or in Ireland since. In Ancient Countryside in England the infrastructure, including the woods, changed little from 1598 to 1844 (see Fig. 177); even if woods were grubbed out they usually left ghost outlines in the hedges. In Planned Countryside the woods usually survived even if the infrastructure did not (see Fig. 179).

On the 1598 map the woods are not defined wood-lots, but patches of wood within fields, rambling inconsequentially from field to field. (None appears to contain a rath.) Was this a landscape of decline, with no longer enough people to hold back the trees? If the Irish woods of the 1650s were in part relatively recent, the result of a century of sword, pestilence and famine, they might have been destroyed without much comment when prosperity returned. Not being demarcated from fields, they would have disappeared without even a ghost remaining.

In 1550 the woods were fully used: industries like Wealden iron and glass, and cities like London and Canterbury, had come to terms with their local woodland. Where did energy and timber come from to supply the huge later expansion of population and industry?

Coal had been used since Roman times. By the Middle Ages most surface coalfields were being worked, and coal was preferred for certain trades. When Edward I was building great castles in North Wales in the 1290s he first needed limekilns: although there was plenty of local wood he shipped coal from distant mines.

Coal was cheaper in labour costs. A miner was reckoned to raise about 250 tons of coal a year, about a ton per working day. A woodcutter would find it difficult to produce a ton of made-up wood a day from the tree to the roadside: this would be equivalent in heat produced to at most half a ton of coal, in practice less because of the water content of wood. The difference in labour cost paid for long-distance transport. Coal was cheap even at a distance.

In 1550 the population of London was about 90,000. Most of their domestic and industrial fuel came from the ring of woodland 20–40 miles (30–60 kilometres) out. They were using about one-fifth of a ton of coal per head per year, shipped 300 miles (500 kilometres) from northeast England. Fifty years later, the population had roughly doubled, and so had the fuel consumption. The woods were producing roughly the same, but coal consumption had risen to about three-quarters of a ton per head per year.

England has an area of 30 million acres (12 million ha). In 1700 about 1.8 million acres (730,000 ha) were woodland. If those woods yielded 2 tons of dry biomass per acre per year, half of which was used for fuel and half for timber and other purposes, this would make 1.8 million tons of dry wood fuel per year, equivalent to 0.9 million tons of coal per year. The actual coal production was at least 1.2 million tons per year. The calorific value of the coal consumed in England had overtaken that of wood in the mid-to late seventeenth century – 200 years before the United States reached that point. By 1815 England was digging 22.6 million tons of coal a year, more (it is said) than all the rest of the world, equivalent to nearly 50 million tons of wood, the annual production of woodland as big as the whole of Great Britain.³³

As E.J.T. Collins points out, industrial use of wood fuel in Britain declined uniquely early. Charcoal iron production in England peaked at about 25,000 tons a year in the 1750s, but in the United States rose to 680,000 tons in 1890, and in the one state of Michigan to 300,000 tons in 1907. This is to do with England having more expensive labour, worse ores, and competition from coal.³⁴

The thesis that woods were destroyed by heavy industries cannot be sustained. On the contrary, wherever there remained a big concentration of woodland, there is an industrial or urban use to account for its preservation. It was the ‘unexploited’ woods that disappeared from the map.

Industries, however, are liable to sudden death through technological change or foreign competition, leaving their woods unemployed. In Cornwall the woods, when the tinners had finished with them, reverted to domestic use, leaving charcoal-hearths (p.166ff) as witnesses to their industrial phase.

Disused industrial woods could pass to other industries. Northern and western coppiced oakwoods were taken over by an expansion in leather-tanning, which used bark and could be combined with other industries using the wood. Others produced pit-props for holding up the roofs of coal mines. The Chiltern woods, as the London market for billets and charcoal declined, were gradually taken over by a mechanised furniture industry. So thoroughly were they converted from coppice and pollarded wood-pasture to timber production that by the nineteenth century this was regarded as the normal state of a beechwood. The rise of shipbuilding would have found a use for the oaks no longer wanted through the decline of timber-framed building.

Specialised underwood trades expanded, especially in southern England, to take advantage of increasing markets for mass-produced hop-poles, barrel-hoops, and other industrial, agricultural and domestic artefacts; this was probably related to the replacement of mixed underwood by chestnut. Indeed, Collins has termed this ‘the golden age of English woodlands’.

The decline of industries left woodland open to destruction. Coed Glyn Cynon was still very much alive in the 1810s, when again there were complaints that its valleys had been ‘stripped of their grown timber’. After the iron industry had died, it seems to have produced pit-props and then passed to modern forestry, which lasted for only one generation of planted trees. On the Ordnance Survey of c.1870 the South Welsh valleys were still one of the biggest wooded areas in the British Isles; by 1950 the woods had faded away, mostly into moorland. In Kent and Sussex, although there is still plenty of woodland left, much was grubbed out in the nineteenth century, the golden age of hop growing. Elsewhere, woods were saved by agricultural recession and fell into disuse or were used as pheasant shoots, until the great onslaught of the 1950s.

Woodland is not wildwood. What wildwood was like in the pre-Neolithic (Peterken’s ‘past-natural’ state) is open to interpretation. What it would be like by now had human activity remained at a Mesolithic level (the ‘present-natural’ state) is open to conjecture. What present woodland would develop into in future if human activity were to be withdrawn (the ‘future-natural’ state) is open to hypothesis.

Fragmenting continuous forest (on the Tansley model) or stopping the dynamics of groves (on the Vera model).
Instituting coppicing and the production of small timber trees.
Giving woods boundaries, ownerships and names.
Withdrawing browsing.
Selecting guilds of plants that take advantage of cyclical changes in shade.

These are not recent events; woods have had an appreciable fraction of the Holocene in which to come to terms with them. Some aspects, such as the removal of upstanding old trees and large deadwood (apart from boundary pollards) have presumably reduced the fauna and flora of the wildwood, though wood-pasture gives an alternative means of survival. However, woodmanship is an ecological factor in its own right, just as mowing is the defining factor in meadows.

Coppicing has produced several guilds of plants that respond to it (p.211ff). Where did these come from? What were they doing before people invented axes? Neither Vera’s nor Tansley’s theory of wildwood seems to provide for large temporary openings every 7 to 30 years.

Fire, where it is possible, stimulates the germination of buried-seed plants, but woods with coppicing floras tend not to be combustible, least of all in England. Alternative natural processes are avalanches, windblow, beaver activity and ice-storms; although some of these can stimulate buried-seed plants like coppicing,³⁵ they are too sporadic in time or too limited in space to be a prototype of coppicing. Are trees and woodland plants still conditioned to some factor from previous interglacials that no longer operates?

fn1 There are other loose ends in the calculation, such as conversion of carbon into limestone in the sea or into peat in bogs, and the expansion of the Sahara desert (which seems to be independent of human action).

fn2 Neither mere has sediments suitable for radiocarbon dating, but the Roman period can be identified with certainty because it comes just before the first appearance of pollen of Cannabis, a specifically Anglo-Saxon crop.