Woodlands

Contrary to what we used to think, the ‘memory’ of the forest is not of just one, two or three centuries, but of at least 2000 years.

Woodland can grow on almost any soil or rock except deep acid peat (though ‘bog firs’ and other fossil trees show that it occurred even there in the past). Woodland soils, however, are not a random selection of all soils. Woods tend to be preserved or formed in particular places, and their soils develop differently from elsewhere.

Woods in Britain and Ireland, to a large degree, are survivors of periods of land hunger in the thirteenth and nineteenth centuries. They are normally not on land that is good for growing trees, but on land that (by the standards of those times) is bad for cultivation through steepness, infertility or poor drainage. Secondary woods have arisen on abandoned settlements or ex-industrial land, heaths and commons, and lately even meadows.

Soil texture: Readers may have been taught that soils are formed by the weathering of underlying rocks. That is so to some extent, but especially in lowland areas soils often derive from separate superficial deposits that may be too thin to appear on a geological map. Thus woods mapped as chalk may have intensely acid soils formed from surface loess, dust that has travelled in the atmosphere. Because woods escape ploughing, they are likely to preserve thin deposits: this is one way in which woodland soils differ from farmland.

Soils contain four main size-classes of material: clay (smaller than 0.001 millimetre), silt (0.002–0.1 millimetre), sand (0.1–2.0 millimetres) and gravel (bigger than 2.0 millimetres), besides coarser cobbles and boulders. These may come from the solid geology, from glacial deposits like boulder-clay, or from blown sand. Silt often comes from loess, wind-blown dust. This mostly results from dust storms around the end of the last glaciation, although it is still being deposited in areas downwind of deserts, like Japan, Texas or New South Wales; Sahara dust rains down every year on Crete, and as I write a cloud of it has even strayed to Britain.

Wet places: A very important factor is whether water accumulates in wet seasons, usually in spring. Waterlogging occurs where there is not enough slope to carry rainwater away, so that stagnant water fills air spaces and deprives the soil of oxygen. It thus encourages plants, such as aspen or oxlip, that tolerate these consequences. It does not have to be continuous or even annual. In Hayley Wood – on a flat hilltop in a low-rainfall part of the country – there has been significant waterlogging in 14 out of the last 42 years, culminating in the supremely wet spring of 2001 (Figs 58, 64).

Flushes are where the surface is irrigated with moving water, especially from springs. Moving water picks up oxygen where it passes cracks or root-holes, and has different effects from stagnant water. Flushes with alder, sometimes bordered with ash, occur in sloping woods in many parts of the country (Fig. 65); on nearly level ground they are associated with the remarkable ‘plateau alder-woods’ of the Bradfield Woods (Suffolk) and in north Norfolk.

Acidity: Soils can be acid or calcareous. Their acidity is measured on a scale of pH, which varies from 3.0 (the most acid soils) to around 8.0 (the most calcareous).

Soils develop with time. Rainwater percolates through the soil and leaches out soluble minerals, at a rate depending on the composition of the soil and its humus content. It may also wash out clay particles, a process known as lessivage.

Woodland soils tend to be acidic, even if they overlie chalk, chalky boulder-clay or limestone; they get more acid towards the top. Part of the acidity may originally have been there in layers of loess or sand, and part has developed over time. Carbon dioxide, which is acid, is drawn out of the atmosphere and synthesised into leaves, which when they fall and rot release organic acids, fizzing away any calcium carbonate that they meet. So a soil profile under bracken, with pH of 3.1 in the bracken litter, may rise to pH 5.6 in the sandy mineral soil and 7.8 in the underlying boulder-clay.

Surface acidification tends to be counteracted (if the soil is not too acid or too waterlogged) by earthworms and moles, and by tree-fall mounds if any (p.164f), which mix the layers. In well-drained parts of Hayley Wood, pH rises from 6.1 at the surface to 7.1 in the boulder-clay.

In ancient woodland, surface acidification may have accumulated over thousands of Holocene winters of rotting leaves. On earthworks such as woodbanks it is observable in miniature. The bank itself usually has an inverted profile, with the calcareous subsoil, dug out of the bottom of the ditch, on the top of the bank. The bank is therefore less acidic than the rest of the wood, but has a surface-acid layer a few inches deep, produced by hundreds of winters of rotting leaves.

Leaf litter and humus: When leaves, twigs, bud-scales etc. fall, one of two things may happen. Earthworms may come out of their burrows at night and drag down leaves to eat below ground: the worms and their gut bacteria disperse the breakdown products among the soil in the form of wormcasts. As Charles Darwin, the great earthworm ecologist, showed, worms are choosy: they take soft tasty leaves like elm first and leave the tough, nasty-tasting oak leaves till last.²

Alternatively, leaves pile up on the surface and are broken down gradually by mites and fungi; they form a stratified humus layer of many years’ leaves in successive states of disintegration. Fungal hyphæ and mites’ excrements give leafmould its distinctive smell.

Leafmould soils are known by the general Danish designation of mor; earthworm soils are called mull. The distinction depends partly on worm activity and partly on the trees. Acid soils are hostile to mull-forming types of earthworm, and the trees that grow on them tend to have unpalatable dead leaves. The transition from mor to mull is typically around pH 4.0, although I have found mor under beech in soils as weakly acid as pH 4.6, and mull under maple at pH 3.5 (on a rare occasion when maple grew on acid soil). Mull is also encouraged by coppicing.

More detailed studies subdivide mull and mor into many different categories. In America, where some forests have great thicknesses of ‘duff’ and leafmould, the matter is complicated by uncertainty as to which earthworms are native and which are European introductions.

Other countries have other ways of disposing of leaf litter. A loose, airy litter structure, as with pines, encourages ground fires. In Australia, many eucalypts produce a litter so unpalatable that fire is the only way of recycling it. In North America, however, sugar-maple leaves pack down into a cardboard-like layer that appears to be a means of suppressing fires.

Fertility – woods versus farmland: Woodland soils are not a simple extension of the soils of the surrounding farmland. They would not have remained as woodland if they had been attractive for agriculture. Early farmers depended on the natural fertility of the soil, since they could not buy fertility in a sack: infertile soils remained as woodland or became heath. In infertile parts of Scandinavia, or in Japan where there was little animal manure, people would gather leaf litter and bushes from the woods to fertilise their fields, but in Britain they have seldom been reduced to such desperate expedients.

Woodland soils have developed independently of farmland. They are not ploughed or limed. They tend to retain their stratification and their top layers, such as thin layers of loess. Ploughing (rather than deforestation) promotes sheet erosion: soils on sloping arable fields creep downhill away from the bottom edges of woods and pile up against the top edges of woods lower on the slope (see Fig. 188).

Ancient woods have escaped the centuries of manuring, followed by two centuries of imported or artificial fertiliser, that farmland has had. They would lose phosphate as successive crops of wood were removed. The phosphate never came back, but ended in gardens, dunghills and churchyards, to nourish plants such as stinging-nettles. Occasional accumulations of phosphate may result from phosphate-bearing rocks, or historic or prehistoric settlements where people have lived and gardened and died, or maybe from a big roost of starlings.

In high-rainfall areas woods may be especially infertile, because nutrients have been washed out by 10,000 years of rain. This is partly why woods have turned into moorland as the Holocene has progressed. It happens to some extent all over the world. Some tropical forest soils, developed over millions of years, are extremely infertile. Many a settler in the tropics has been lured to ruin by the misbelief that the ‘luxuriance of nature’ implies fertile soil.

A growing conservation problem is contamination of woods and other nutrient-poor places by fertiliser dust or fertilised soil blowing off the fields, so that common, vigorous phosphate-dependent plants overrun the historic woodland vegetation. To anyone wanting to imitate an ancient wood a chief obstacle is that an ex-farmland site is too fertile. In very low-nutrient environments such as Atlantic oakwoods, nitrogen deposition may affect bryophytes and lichens on trees.³

Moisture: Drought can cause dieback of trees and even in Britain can affect woodland. Places waterlogged in spring may be droughted in summer: waterlogging prevents the plants from putting down deep enough roots to withstand drought. Drought in dry summers is apparently what stops oxlips in Hayley Wood from being immortal, rendering them liable to extermination by deer.

In other countries drought has more obvious effects on wild vegetation. In the Mediterranean, summer is the dry season, to which many plants are ill adapted because their evolutionary ancestry obliges them to grow in summer. Trees are often lacking, or reduced to the stature of bushes (maquis). Alternatively, trees may be widely spaced to form savanna: widely spaced above ground, but filling the whole below-ground space with their roots, so that they capture rain falling between the tree crowns as well as on them (Fig. 66).

It used to be claimed that the Mediterranean was originally covered in dense forests, and that the present savannas, maquis and garrigues are forest that has been ‘degraded’ by thousands of years of people cultivating, cutting wood, keeping livestock and burning vegetation to create pasture. Important though these have been, there can be little doubt that the limiting factor has been moisture, ever since the climate reached approximately its present state some 5,000 years ago. Moisture depends on rainfall, on whether the soil or bedrock can retain rainfall, and on whether the roots can penetrate the rock to get at the water. Moisture-retaining schist rocks may be more vegetated than limestone in the same climate; karstic limestone (with holes and fissures) may be more vegetated than solid limestone. Trees tend to grow on screes, or along a geological fault, or on the ruin of a Roman farmstead, or over an ancient tomb – features that promote root penetration.⁴

Mediterranean savannas are chiefly in the region of roughly 16 to 24 inches (400 to 600 millimetres) annual rainfall: less rainfall and there is steppe, more rainfall and there is forest. In Texas, where savannas are, again, partly cultural, and where rain falls all the year, they occur in the 16 to 40 inch (400 to 1,000 millimetre) belt. Although much of England falls within these limits, the less hot summers make moisture go further, so that climatic savanna is hardly a possibility – with the possible exception of Staverton Park, an extreme combination of low rainfall and poor moisture retention.

In hard-rock landscapes, trees or woodland tend to be on rock: areas where there is soil are grassland or heath. An English example is Bradgate Park, a Midland outlier of the Highland Zone in which the ancient trees are mostly on rock outcrops (Fig. 67). In South Wales small, irregular woods followed (and occasionally still follow) very closely the rocky patches in the landscape.

In Estremadura (Spain) the dehesa wood-pastures are on shallow soils over granite: areas where clay has accumulated tend to be treeless farmland. In middle Texas woods tend to be in canyons and on limestone outcrops; where clay overlies the limestone, trees are lacking. Although large extents of clay were cultivated by settlers, land-grant surveys show that when they arrived in the nineteenth century the claylands were prairie, the limestone soils were savanna, and the canyons were wooded. In west Texas, savanna on the Davis Mountains ends abruptly at the base of the volcanic rocks and gives way to the stoneless plains of the Chihuahua Desert. I could multiply examples from Crete, Japan and other countries.

This phenomenon – trees on bare rock, grassland on soil – has escaped the attention of ecological writers, who were taught that trees cannot grow without soil. It is more than the mere effect of people cultivating the areas with soil. Rock fissures may have some attraction for tree roots, such as access to deep stores of moisture in a semi-arid environment. Or (in flammable countries like middle Texas) rocky areas may interrupt the spread of prairie fires. It is a reminder that in tree’d grassland the trees are not always the dominant partner.

In most natural woods the trees have been coppiced: lack of coppice stools raises a suspicion that the wood is either recent or much altered by modern forestry.

Cut down a maple-tree, and it will sprout and form a small stool; cut it again after ten years, and it will sprout again and form a bigger stool; after 300 years it will be some 3 feet (1 metre) in diameter. In New England, where coppicing was introduced by English settlers, wood-lots contain stools of Acer rubrum of up to that size. In Old England, stools of A. campestre can be much bigger, resulting from cycles of felling and regrowth over a much longer period.

Such giant stools (Figs 68, 69) are living archaeology, independent evidence of the age and management history of a wood. Their ages can be estimated by looking for the biggest stools of a given species in a wood of known date of origin. In the Baltic islands of Åland, Carl Hæggström found hazel stools more than 6 feet (1.8 metres) in diameter, the oldest of which he estimated – from the date when the land they stand on emerged from the sea – to be 990 years old.⁵ Hazel in nineteenth-century plantations is up to about 2 feet (60 centimetres) in diameter.

Alternatively, annual rings can be counted and measured in a section of the above-ground base of a stool. Giant stools are hollow, and usually have only a shell of wood from which to estimate the rate at which they have expanded. The last few coppicings can be dated by looking for cycles of narrow and wide annual rings. This source of information was all too common in the 1960s and 1970s, when ancient stools were destroyed every day, but is now less easily come by.

As a very rough guide, a coppiced ash stool 4 feet (1.2 metres) in diameter would be 400 years old, and one 8 feet (2.4 metres) across would be 800 years old. On wet or infertile sites stools are likely to expand more slowly than this.

Lime, oak and hazel stools probably grow about as fast as ash; maple a little faster. Sycamore stools must grow faster, for they can be 10 feet (3 metres) in diameter, although the tree has been in Britain for only 450 years. (It was fashionable in the seventeenth century, and many sycamores planted then, or their children, must still be alive.) Chestnut is also a fast grower: in Felbrigg Great Wood (Norfolk), where it was first planted in the 1670s, some stools are 7 feet (2 metres) across. The biggest chestnut stools that I know are up to 16 feet (5 metres) across in Holbrook Park (southeast Suffolk) and Stour Wood (northeast Essex).

Very large coppice stools grow out into rings, raising the question: Is this one individual, or an amalgamation of two or more? This can be decided if all the stems on a stool are one clone, sharing some individual peculiarity of bark, branching habit, time of leaf-opening or leaf-fall, etc. In ash the first-year shoots after coppicing contain a red pigment that distinguishes them as individuals by the intensity and colour of the pigment and its distribution – whether it is diffused or concentrated around lenticels. Usually all the stems of a big stool share the peculiarity and differ from neighbouring stools (Figs 70, 71) : they are thus parts of the same individual. Occasionally a big stool shows two peculiarities divided by a line through the middle: it thus consists of two individuals close together.

Lime shows similar peculiarities, but as with ash some of them are visible for only a short time in the year or in the coppice cycle. Lime, unlike ash, might reproduce vegetatively in two ways. Some limes are self-layering: weak stems bend over to the ground and take root at the tip, which should generate circular clonal patches. Alternatively, limes can blow down in storms and take root along the trunk, creating linear stands of genetically identical stems (Fig. 72). Both are probably uncommon: in my experience identity extends to no more than three adjacent stools.^fn1

This inference is confirmed by David Morfitt’s detailed observations in Piles Coppice (see below). He searched for neighbouring stools sharing some peculiarity of habit, pigmentation, time of leafing or leaf-fall, etc. The results were largely negative. Many stools are distinct from all their neighbours; possible clonal patches seldom extended to more than four stools in a row up to 120 feet (35 metres) apart. Some of these, inevitably, will be false matches, for it is easy to prove that two individuals are genetically different, but (even with DNA analysis) less easy to prove that they are identical. Vegetative reproduction has been less important for these limes than propagation by seed. If, as is likely, the limes were coppiced from the Middle Ages until the twentieth century they have had little opportunity to blow down.

Giant stools occur in most other countries with a long coppicing tradition, such as France, Norway, Sweden, Italy, Hungary⁶ and Greece; also in Japan, though less commonly than here (Figs 73, 74). In many countries one must beware of coppicing as a response to fire, or of self-coppicing, as in Japanese or American limes and magnolias. The live-oak motts of inland Texas (Fig. 42) are, in effect, super-stools.

The biggest coppice stool that I know of is the now famous lime (Tilia cordata) in Westonbirt Arboretum, Gloucestershire, some 50 feet (15 metres) in diameter; it is a relic of the medieval Silk Wood (Fig. 75). It all has the same time of leafing, and I am told has the same DNA. Nearly as large is a lime ring in Gosling’s Corner,^fn2 a surviving fragment of the great Langton Wood near Wragby, Lincolnshire. Such stools have a claim to be the oldest trees in Britain and the last surviving trees of the wildwood: they could well have originated by self-coppicing.

Oaks: Oak has been symbiotic with people. For over a thousand years, it was the commonest and most expensive timber tree. Woodmen, finding a young oak, did not cut it with the underwood, but let it grow to a timber tree. Although oak will grow in almost any environment in Britain except on thin chalk soils, this gave it a competitive advantage. In the eighteenth century rising prices of oak bark and oak timber encouraged landowners to increase the proportion of oak. Even after the bottom dropped out of both markets in the 1860s, many owners still encouraged oak.

Then came the Oak Change, around 1900: both species of oak largely stopped replacing themselves within existing woods (p.335f). With a few exceptions such as the Bradfield Woods (Fig. 76), most woods are now in an unhistorical state as regards their standard oaks. Instead of a mixture of ages, with young trees predominating, there are now only oaks of 100 years old or more.

Standard trees in woods that have escaped replanting now tend to be in one of three states:

Most of the oaks were felled between 1914 and 1945 and have not been replaced (except for sporadic regrowth from stumps). The stumps remain and give a rough indication of the date of felling (p.165f). About one-third of ancient woods are in such a state, mainly the bigger ones
The oaks are still a wild population, but few or none are younger than a century. The smaller oaks now are roughly as big as the biggest oaks would have been in the Middle Ages. Stumps indicate whether there has been any felling since 1900. The oldest oaks are seldom much over 200, although there are exceptions (Ken Wood in Hampstead, London, has some from the 1680s).
The oaks are not a wild population, but result from replacement of wild-type oaks with nursery-grown trees of what were regarded at the time as ‘good’ genetic stock, as was the practice between about 1820 and 1920. These oaks have a degree of uniformity that distinguishes them from wild oaks (for details see Chapter 13). This has happened in a large minority of woods (often only part of a wood), probably more in the Highland Zone (Chapter 15).

In a rational world, timber trees would not be scattered throughout a wood, but would all be together in one part. This arrangement should result in better timber and better underwood, but was not often practised. The oakwoods of west Cornwall (and I am told south Devon) are divided into a timber part, located in a sheltered ravine, and a wood-producing part occupying the more windy slopes. In some of the woods on the Blean (east Kent) timber oaks are concentrated into particular parts of the wood, delimited by rides, the rest of the wood often being chestnut coppice.

Other species: The beechwoods of the Chilterns, consisting entirely of timber trees, appear to be a nineteenth-century development from coppice and wood-pasture (p.283f), but have still some connection with natural woodland in that most of the beeches are wild-type.

Other timber trees for which there was a market were elm and ash. Most of the elm came from non-woodland sources, but surviving elmwoods have standard trees as well as outgrown underwood. Ashes as timber trees are rarely more than 150 years old. Those that are older, as in Hayley Wood, go hollow and begin to take on the character of ancient trees.

Other native species were historically regarded as underwood, and rarely allowed to reach timber size before the twentieth century. Miller Christy, nearly a century ago, remarked of hornbeam: ‘a mature wild-grown example in its natural condition is very seldom seen’, but such are now not uncommon. The same is true of maple, lime, cherry, crab and even service. Woodmen have allowed these to grow up, as a curiosity or in the hope that a market for their timber might arise. Latterly it has become the custom, when coppicing in nature reserves, to promote these other species instead of the missing young oaks.

Ancient woods are not the place to look for ancient trees, apart from coppice stools. Indeed the presence of ancient trees, unless they are boundary pollards, indicates that the wood is not ancient, but has grown up around freestanding trees (infilled savanna, p.125f). However, some standard trees in woods are on the way to becoming the old oaks and ashes of the future: they are acquiring conservation importance because they are difficult to replace.

No visitor to the Australian tropics can forget the massive ‘fern-gardens’ high in the crowns of rainforest trees, or the climbing rattan palms that hang down from far above, armed with razor-wire prickles whose least touch draws blood. Most of the action in such forests, and the diversity of species, happens out of sight high in the trees. As memorable are the gnarled wisterias of Japanese sacred groves, their empty coils enveloping the ghosts of the trees that once they climbed, or the mysterious little ferns that proliferate on the soft bark of the giant kusonoki tree. In the southeastern United States mighty vines loop from tree to tree and even, enigmatically, across roads.

Britain has few of these. We have five woody climbers: ivy, adhering by aerial roots; honeysuckle, twining; clematis, using its leaf-stalks as tendrils; woody nightshade, scrambling informally; and dog-rose, hooking itself up, rattan-wise, on prickly flagellar shoots. Epiphytic vascular plants – rooted on the tree rather than climbing up it from the ground – are mainly in western oakwoods, especially the polypody fern on big horizontal boughs (Fig. 77), or wood-sorrel in rot-holes.

Pollards have many more epiphytes, even, on occasion, including other trees. A few big pollard willows, like a few big coppice stools, are genetically more than one tree. A second willow may start from seed in the crown of the first tree, root through the central cavity into the ground, and mingle its tissues with those of the host tree.

Rural Warwickshire is traditionally divided into two halves. To the northwest is Arden, an Ancient Countryside of hamlets, ancient hedges, lanes, holloways and the ghosts of former heathland, often mistaken for Shakespeare’s Forest of Arden.^fn3 To the southeast is Feldon, a Planned Countryside of villages, Enclosure-Act roads, hawthorn hedges and straight lines. In Anglo-Saxon times Arden had a huge extent of woodland, much of which had already gone by the thirteenth century; the monumental work of Sarah Wager describes – as far as documents reveal – what happened to each of the woods.⁷

On Arden’s eastern fringes, extensive ancient woodland lingered until the twentieth century’s vogue for grubbing and replanting; much of the replanting was unsuccessful and the ancient woodland has reasserted itself. This area has been studied by David Morfitt. Around Coventry, Domesday Book records a big concentration of woodland, which probably already contrasted with Dunsmore Heath to the east – an example of the ancient association between woodland and heath.

On the very edge of Arden is Piles Coppice in Binley (52 acres/21 ha), where Dr Morfitt’s studies were concentrated (Fig. 198).⁸ It is now crammed between a railway, a motorway, electricity pylons, a housing estate and a conifer plantation on the site of an ancient wood; but like many urbanised woods it is in surprisingly good condition. It has belonged to the Woodland Trust since 1987.

Piles Coppice sits on a ridge between two shallow valleys; it has sandy-silty soils probably containing loess, and quite strongly acidic (pH 3.5–4.6). It is largely surrounded by a woodbank with ditch on the outside, except on the southeast where the earthwork is poorly preserved and faces the other way, showing that it belonged to the adjacent territory of Brandon. It has an impressive spring flora.

The characteristic trees are lime and sessile oak. Small-leaved lime is dominant on the sloping parts of the wood, as stately coppice stools long outgrown to nearly 100 feet (30 metres) high; they are mixed in the usual way with standard oak-trees. Dr Morfitt found that the stools are of various diameters and ages, the largest some 11 feet (3½ metres) in diameter and evidently many centuries old. The commonest size, around 5 feet (1½ metres), indicates that there has been some more recent replacement of lime. In each part of the wood that he examined, giant and big stools occur together.

Lime is infrequent in the surviving Arden woods, but there is an area of limewood in the interior of the next ancient wood but one, Birchley Wood; this too has big stools in an embanked wood that is well documented back to c.1150. (The name, anciently Burtleia, has nothing to do with birch.)

The oaks in Piles Coppice comprise Quercus robur, Q. petræa and intermediates which are usually regarded as hybrids.⁹ This was confirmed by Dr Morfitt using leaf measurements for the two species. There are both timber trees, interspersed among lime and hazel coppice, and coppice stools. Timber oaks, mainly in the lime area, were of both species with many intermediates. Coppice stools were overwhelmingly sessile oak, varying in size from those only once cut to ancient stools 8 feet (2½ metres) across.

Binley, anciently Bilney, is fairly well documented. A map of 1746 shows ‘Piles Coppice’ almost exactly as it is now, but before this there is a gap of nearly five centuries, bridged by a complex and confusing series of charters. Dr Morfitt identifies Piles with a private wood 4 × 2 ‘furlongs’, which Domesday Book traced back to before 1066; the area and shape (½ × ¼ mile, 0.8 × 0.4 kilometres) are about right. At some time after 1150 it was bestowed on the Cistercian monks of Combe Abbey, who had it for the rest of the Middle Ages.

A wood called Coppice or Copse is likely to have been so-named to contrast with a nearby non-coppice wood, probably a wood-pasture. Near Piles Coppice was Binley Common Wood, now mostly swallowed up in the new town of Binley Woods. (Housing developments tend to be named after what they destroy.¹⁰) On the south, the anomalous woodbank was part of the compact curved outline of Brandon Old Park. In 1279 what appears to be this wood is recorded as ‘42 acres of land¹¹ of which 2 acres are included in the park of Brandon by payment of 2 shillings and a buck per annum.’ This would have been a common transaction: a rich man planning a park would pay large money for projecting bits of his neighbour’s land, in order to square off the outline of the park and make it economical to fence.

Herbaceous plants, like coppice stools, are a defining feature of a wood. The National Vegetation Classification is, to a large extent, a classification of woodland ground vegetation as it was in the 1980s.

Much is known about the range of environments in which specific plants grow, or (at least as important) in which their competitors do not grow.¹² Environmental factors vary in several independent dimensions: soil texture; soil acidity; waterlogging; flushing; degree of shade; time elapsed since last felling. Trees also act as an environmental factor for the herbaceous plants, though the correlation is not particularly close (Chapter 14).

Herbaceous plants (and tree seedlings and ground-living bryophytes) may form guilds, assemblages of species correlated with a particular environment of soil, shade or state of coppicing. Alternatively, one species may become so strongly dominant as to squeeze out all others. Dog’s-mercury is very competitive, being clonal and not needing to establish from seed, and also because it produces a dense canopy of nearly evergreen leaves. Where the environment is ideal for it, it flourishes and squeezes out all other species, even tree seedlings. In less-than-ideal circumstances it coexists with other species as one of a guild. Plants like primrose and orchids, or non-gregarious trees like crab, always appear within guilds.

Why do guilds exist? Why does one best-adapted species – dog’s-mercury, bramble, bracken – not always out-compete the others and take over the site? (Believers in the ‘Tragedy of the Commons’ assert that this is what happens in human affairs.) Are guilds chance groupings of plants that happen to grow in the same environment, or are they plant ‘communities’ in a real sense, with mechanisms of integration between species? These questions have been debated for 70 years.¹³

Herbaceous plants are related to particular soil types more closely than most trees. A classic example of such relations is in the west Cambridgeshire woods, where there is (or was) a series of plant communities associated with waterlogging (Figs 78, 79, 80) :

These zones were studied in the 1940s by B.A. Abeywickrama, and after him by successive students of Alex Watt, David Coombe and Donald Pigott – combining field, garden and laboratory experiments. It was first established that competition was involved: oxlip and bluebell would grow well in mercury territory provided the mercury was suppressed. Next it was shown that waterlogging determined the competition; this was done by raising or lowering plots of vegetation, which adjusted themselves to the new drainage status. Waterlogging might be expected to act via poor aeration, lack of oxygen in the soil or accumulation of carbon dioxide, but turned out in practice to work through converting iron compounds from the ferric to the ferrous state. Ferric iron is harmless, but ferrous is toxic to plants, whose sensitivity closely corresponds to their position in the waterlogging scale. Oxlip is very tolerant and mercury very sensitive: oxlip grows in the worst-drained places, not because it prefers bad drainage, but because it cannot compete with mercury.¹⁴

In runs of dry winters, mercury, which is clonal, advances into hollows, and dies back when these are flooded. To a lesser degree waterlogging affects soil acidity: soils are more acid in areas liable to waterlogging.

Flushing attracts a different guild of plants, including ramsons Allium ursinum, golden saxifrage, Equisetum telmateja and Carex strigosa.

Soil acidity is another determinant of ground vegetation. All the plant communities mentioned above are relatively calcareous, though bluebell can extend on to very acid soils. Woodland plant communities on strongly acidic sites involve bracken, some brambles, the grasses Molinia cærulea and Deschampsia flexuosa, or Sphagnum. The National Vegetation Classification assigns most ancient woods to variants of just two groups, W8 (calcareous or less acid) and W10 (strongly acid), the line being drawn around pH 4.8.

Places with accumulations of phosphate, natural or archaeological, tend to have a distinctive suite of plants, especially nettles, but also goosegrass, ground-ivy and the grass Poa trivialis.

The degree of shade under different trees varies rather little in winter, but much more in summer, when most of the light has to pass through leaves. Ash (for example) casts a light shade, whereas lime is heavily shading. Another factor is the time when the tree canopy comes into leaf: early-leafing trees can deprive the ground vegetation of three weeks’ extra light at the brightest time of year.¹⁵

Some woodland plants avoid shade; others tolerate it, responding to lack of light by producing thinner, more spread-out leaves that capture photons more efficiently. Many, such as meadowsweet or Deschampsia cespitosa, die out under dense shade (e.g. hornbeam); they persist in moderate shade, but flower only under light shade (e.g. ash). A few, especially parasitic orchids and some bryophytes, appear to prefer the densest shade. Shade maintains some guilds by keeping out aggressive, light-demanding species, especially grasses.

Woodland grassland plants, which need to see the sun through permanent gaps in the tree canopy, include valerian, devil’s-bit scabious and some sedges.

Woods were not always dark and gloomy as they often are now. When they were regularly felled a profusion of plants responded each time. These would build up reserves in the first year after felling, flower in the second (less often the third) year, and gradually decline as the underwood grew up and re-established the shade.

Felling a wood (but leaving a scatter of timber trees) typically increases light in summer by at least 20-fold; increases light in spring (before the trees come into leaf) three- or four-fold; and in many woods extends the period of spring light by about three weeks, because the remaining shade comes from standard trees that come into leaf later than the underwood.

Spring-leafing persistent perennials. These dozen or so species include some that are common and important: bluebell, primrose (Fig. 81), anemone, lesser celandine, early purple orchid. They are there all the time, but rejuvenate themselves every time the wood is felled. They respond strongly to coppicing, even though they do most of their leafing in spring when the effect of removing the shade is less than in summer. The shade phase is necessary to them because it keeps down stronger-growing competitors that require continuous light.
Summer-leafing persistent perennials. This larger group includes violets, meadowsweet, water-avens and several grasses such as Deschampsia cespitosa and Melica uniflora.
Buried-seed plants. This large group (at least 100 species) are plants invisible between coppicings. They emerge from a seed bank laid down the previous time the wood was felled, spring to life, flower in the second year, renew the seed bank, and return to dormancy. Foxglove in woods on acid soil can appear in millions where not one was visible in shade. This mysterious capacity is widespread: one-third of the entire flora of the wood may be invisible if there has been no recent felling. Many rushes, speedwells, St John’s-worts, brambles, wood-spurges (Fig. 83), ragged robin and even heather behave in this way.
Buried-seed plants discriminate between individual woods. The ‘signature plant’ of Hayley Wood is ragged robin, of the Bradfield Woods wood-spurge Euphorbia amygdaloides, of Hempstead Wood (northwest Essex) wood forget-me not Myosotis sylvatica, of Chalkney Wood red campion and wild raspberry, of Groton Wood trailing St John’s-wort Hypericum humifusum. Although these are plants with definite ecological ranges, it is difficult to account for this local abundance in terms of environmental factors. The development of a distinctive set of coppicing plants for each separate wood may be one of the processes in the change from wildwood to individual woods.
I leave as an exercise for the reader this question: how do buried seeds know that the wood has been felled and it is time to germinate? Being buried, they can hardly see the extra light. It could be some microclimatic effect, such as higher temperature in a coppiced area; but if so, why does not the whole seed bank germinate in vain every time there is a season a few degrees warmer than average? Or are a few per cent of the seeds programmed to germinate every year – but only if there is extra light do they grow into plants? (Without extra light they succumb to something – let us call it slugs – before they get big enough to be noticed.)
Mobile species. These, especially willowherbs and thistles, have an efficient dispersal mechanism and move around the wood, chasing newly felled areas.
Tree seedlings. Trees such as ash and birch, and formerly oak and hazel, rely on coppicing to progress from the seedling to a young tree.
Aquatics. Aquatic plants, such as brooklime speedwell and water violet, occur in woodland ponds, but are visible only after felling. Water violet probably spends the interval as vegetative buds buried in the mud.
Casuals. Many plants, such as hemlock, mullein, borage or the occasional stray cereal, do not persist in woods, but come haphazardly into felled areas from outside.
Unresponsive. A dozen species appear to be unaffected or even set back by coppicing. These include ramsons, adder’s-tongue fern and herb paris. Dog’s-mercury is damaged by coppicing, which may benefit other plants by breaking (for a time) its competitive monopoly in woods that favour it.

Much recent ‘conservation’ coppicing is akin to silvicultural thinning, leaving (by historical standards) too many trees standing and providing more continuous shade. Where there are deer this weakens the regrowth; where there are not it favours brambles over more specialised coppicing plants. Thinning versus coppicing, when compared experimentally in a French hornbeam–oakwood, produced strikingly different results. Thinning produced a ‘spectacular spread’ of brambles, and also grasses and sedges. Other guilds of coppicing plants were fully developed only after coppicing. Although thinning was presented as a ‘close-to-nature’ practice, it did not replicate the way that woods actually work.¹⁷

The appearing of coppicing plants is one of the changes that transformed wildwood into managed woodland. Where did they come from? They are not pioneer plants of open ground, nor are they the same as plants of permanent open areas (woodland grassland plants), nor (apart from casuals) are they universal weeds. They are by no means related to the sort of woodland that can be burnt (p.44f). In wildwood, natural treefall, from storm or avalanche, seems too rare an event at any one place to give rise to them. However, part of the Tatra Mountains, Slovakia, where blowdowns repeatedly occur (p.19f), has something closely resembling the coppice-plant guilds of England, including oxlip, water-avens, raspberry, hemp-nettle and other species (Fig. 82). Whether evolution was involved is discussed in the final chapter.

The biggest immediate threat to woodland is browsing animals: deer, sheep or feral goats (Chapter 22). Large herbivores subtract much of the woodland ground vegetation, replacing it with browsing-adapted plants, especially grasses. They render coppicing impracticable. They convert a woodland ecosystem into trees plus grass, with no long-term future for the trees. This has been shown by numerous exclosure experiments. Hayley Wood was established in 1962 especially to protect oxlip; oxlip has declined ever since except where protected from fallow deer. Since seven-eighths of the wood was deer-fenced in 2001–2 there has been a notable recovery except in the one-eighth left unfenced.¹⁸

There is one set of woods that has been exposed to intermittent browsing for centuries, namely compartmental wood-pastures, especially Forests such as Hatfield or Wychwood Forest.

These had a specific grazing regime, typically ungrazed in the first half of each coppice cycle (p.121f) and grazed in the latter half. Recent observations on grazing in woods are mainly concerned with deer or sheep, but in historic wood-pastures there were usually all the domestic animals, and often deer too. If the system worked as it was supposed to do, trees would be protected from browsing in the vulnerable first few years after felling. At the same time herbaceous plants would be freed from browsing as well as shade. Shade-bearing herbs, if palatable, would be exposed to attack in the second half of the cycle; buried-seed plants and mobile plants would be exempt, being no longer visible.

Historical surveys rarely identify underwood, but an exception is the 1564–5 surveys of Rockingham Forest, analysed by G.F. Peterken, which report thorn (hawthorn or blackthorn) in nearly every one of 106 coppices and maple in the great majority; ash was notably rare.

Pasturage in Rockingham continued well into the nineteenth century. J.A. Best has compared trees and ground vegetation in the formerly grazed Rockingham coppices with woods that have no history of intermittent grazing. I shall put this study beside my own records of the 12 surviving coppices in Hatfield Forest, compared with the two purlieu woods close to the Forest and with 12 similar woods outside the Forest and within 10 miles (6 kilometres) of it (Table 14). All sets of woods are on predominantly calcareous substrates, although the soils may be acidic.¹⁹

The Hatfield Forest woods have a history of periodic browsing by deer, cattle and sheep; the purlieu woods, in theory, were accessible to deer only; the other woods should not have a browsing history. In recent decades the separation has to some extent broken down as deer have increased outside the Forest: the records date mainly from 20–25 years ago when this was less of a problem.

Trees favoured by periodic browsing: Maple and hawthorn have long been more abundant in Forest woods. This was found by Best in Rockingham and appears strongly in the Hatfield records. In Hatfield, however, most of the hawthorn is a recent invasion: in the 1920s the wood-fences were allowed to decay and let in cattle, which devoured the underwood, creating gaps which hawthorn later filled.

Blackthorn has been favoured by intermittent browsing in both Rockingham and Hatfield. (It is significant in Rockingham as the food-plant of the rare black hairstreak butterfly.) As with the previous species, this is a matter of abundance rather than presence or absence. The purlieu woods follow the non-Forest woods. Unlike Best, I do not find crab-apple to be commoner with browsing.

Aspen is known in Hayley Wood to be distasteful to deer and to be encouraged by coppicing plus deer.²⁰ In Hatfield Forest it is a clonal and rather uncommon tree, and is limited by requiring waterlogged ground, standing out in wet places where deer have eliminated everything else.

Oak as underwood is a special and anomalous feature of some of the Hatfield coppices: it is not on the very acid, infertile soils where it occurs elsewhere in Lowland England. I have argued that this relates to a breakdown of the separation of coppices and plains, perhaps in the seventeenth century, which allowed cattle to get into the young underwood. Oak, being then prolific and somewhat resistant to browsing, filled the resulting gaps, creating the unusual situation that there were so many oaks that it became worthwhile to treat some of them as underwood rather than timber.²¹

Elm is palatable to all livestock; continuous exposure to deer is the main reason for woodland elm failing to recover from recent Elm Disease. It is surprising to find elm systematically more abundant in Hatfield Forest than in non-Forest woods. However, browsing in the latter part of each coppice cycle might well do it no harm. After each felling, suckers would arise from the roots and in six years would be big enough not to be much harmed when the animals were let in.

Trees disfavoured by periodic browsing: Ash is the only tree which Best or I found to be disfavoured compared to other trees. It is very palatable; but deer normally do not kill ash saplings, but reduce them to gnarled stubs, which wait until reduced shade and reduced browsing let them get away. In Hayley Wood a period of coppicing and severe browsing in the 1960s killed many of the stools. Part of the area was fenced in 1980, and within 14 years was crowded with pole-sized ashes with gnarled bases, filling the gaps.²² This might be taken to indicate that intermittent browsing favours ash (Fig. 85). This, however, is contrary to Best’s experience in Rockingham and mine in Hatfield Forest. Were the seven-year fencing phases too short to sustain ash? Or were cattle involved, which destroyed ash more effectively than deer?

Herbaceous plants favoured by periodic browsing: Best remarks on the ‘grassy’ appearance of Rockingham coppices subjected to grazing, especially due to Deschampsia cespitosa. This grass is more abundant in Hatfield Forest than in purlieu or non-Forest woods, often dominant in coppiced areas; its history in Forest coppices appears to go back to the seventeenth century. Brachypodium sylvaticum adds to the grassy appearance. Although European grasses are adapted to grazing, woodland grasses are either tough and relatively unpalatable, like Deschampsia, or feeble and not worth eating, like Brachypodium.