Woodlands

… From 1940 to 1945 I was concerned with fire-fighting arrangements over many thousands of acres of forest in the South of England … these were subject to unusual risks from large-scale military training and aircraft, and … German incendiary bombs … any kind of vegetation that could be set alight, was set alight, and had of course to be tackled by fire-fighters. Broadleaved woodland of any kind simply refused to burn at any time, although fires in coniferous plantations, and among heather and gorse, were serious and frequent.

Who understands tree roots? In countries like Greece that are fond of the bulldozer, roots are exposed to view in road-cuts; but until 1987 few English people understood what a tree’s root system looked like; some thought roots went nearly as far below ground as stems above it.

As the great storms of 1987 and 1990 showed, most trees in England are shallow rooted (Fig. 18).¹ It may be argued that deep-rooted trees were never uprooted, but anyone digging holes in a wood seldom meets roots more than 3 feet (1 metre) down. A giant beech can have a root-plate only a few inches deep, much less than the diameter of the trunk. A chestnut stool on clay or loess, 8 feet (2.5 metres) in diameter, has all its visible roots in 9 inches (23 centimetres) or so. Trees do not necessarily have deeper roots than herbaceous plants: I have excavated barley roots down to at least 8 feet (2.5 metres).

Is it true that trees in Britain have taproots? An oak, on germination, produces a vigorous vertical root that nails the acorn to the ground. Maybe people thought this went on developing for many years, but oaks uprooted in storms show that this is not so: the taproot goes down a foot or two (30 or 60 centimetres) and is then superseded. David Maylam tells me of oaks forcibly uprooted that proved to have taproots, but it is not clear in what circumstances these are developed. In Greece deciduous oaks generally have shallow roots like English oaks, but evergreen oaks are deep rooted, especially if rooting into rock fissures.

A tree’s roots have three functions: to hold the tree up; to supply it with water; and to supply it with minerals. (Trees usually delegate the last to mycorrhizal fungi.) In England, generally, the holding-up function is the critical one. In the great storms many millions of trees were uprooted. To most people’s surprise they stayed alive; many of them lived through the great droughts of 1989 and 1990, and some flourished better than many trees that had stayed upright (Fig. 19). The same happened after other hurricanes and in other countries. James Dickson points out that woods around Glasgow are full of living oaks and larches overturned in the great storm of 1968. Most trees lived (and are still alive if not shaded) if one-quarter or one-sixth of their root system remained in the ground.

The inference, that trees have four to six times as much root as is needed to supply them with water in an ordinary summer, explains some anomalies. Urban trees often have their roots severed by builders’ or cable-layers’ trenches: this can hardly be good for the tree, but it rarely suffers obvious above-ground effects. Ploughing around a farmland tree must play havoc with its shallow roots, yet the consequences are less severe than would be expected.

Trees should thus be able to grow so close together that they fail to stand up before they suffer unduly from root competition. They do: in the 1987 storm, trees in plantations were more often uprooted than in natural woods, in natural woods more often than freestanding trees, and in the interior of a wood or plantation more often than on the edges, including the windward edge (see Fig. 9). Evidently isolated trees can develop a bigger root system, which holds a greater weight of soil to counterbalance the wind pressure, than their sisters whose roots are restricted by neighbouring trees. A marginal tree develops half an unrestricted root system. Trees in plantations are even more crowded than in woods, partly because foresters earn their living by growing stems, not roots, and accept wind-blow as a normal risk of business, but also because they plant trees close together and often forget to thin them. Crowding turned out to be the most significant factor in uprooting – and also in wind-breakage, but that is another story.²

In other climates the balance between root functions is different. In the Mediterranean, summers are fiercely dry, and most trees have not adapted to drought by losing their leaves as they do in the seasonally dry tropics. Instead they have deep roots and grow widely spaced. In a high wind, most trees break before they uproot.

There are thousands of species of specialised woodland fungi. Many conspicuous toadstools go with particular trees. The familiar fly agaric (is it still familiar?) is always near either pine or birch: its mycelium – the permanent, usually unseen vegetative network of microscopic hyphæ – is a partner with the roots of these particular trees.

Most land plants are dual organisms. Attached to their roots is a fungus, whose hyphæ are thinner and more richly branched than the root itself; they invade more soil than is directly accessible to the roots. The host plant supplies the fungus with the carbon needed to make its hyphæ. The fungus does much of the job which schoolchildren used to be taught was done by the root hairs. It supplies the plant with nitrogen, phosphorus and other nutrients, and sometimes water too; it can even defend its host against competition from non-mycorrhizal neighbours. Neither functions well without the other; seedlings use their seed reserves to make contact with the fungus, and die if they fail to find a partner.

Lichens, too, are dual organisms – a combination of a fungus and a green or blue-green alga. (So is the human body, as anyone knows who has taken an antibiotic that kills off, temporarily, the gut bacteria.) Mosses and certain families of herbaceous plants lack mycorrhizas, but most liverworts have them.³

Mycorrhizas have been known for more than a century, but their full significance has only recently been elucidated by such scientists as Professor D.J. Read.⁴ Most plants have endotrophic or vesicular-arbuscular mycorrhizas. The fungus forms micro-trees inside the cells of the root, and sends out hyphæ that branch far into the soil. These simple fungi, which lack visible fruit-bodies, form a special phylum (Glomeromycetes). It is usually said that the fungal species are few and not particular as to their hosts, but this may be due to incomplete knowledge. Their hyphæ secrete glomalin into the soil, a protein that, like humus from worm casts, stabilises and aerates the soil. This plant-fungus symbiosis goes back to the remote geological origins of land plants.⁵

The second kind of mycorrhiza is ectotrophic: the fungus forms a sheath investing the root, from which strands of fungal tissue invade the surrounding soil. Host plants include trees such as pines, oaks, beeches, birches and hazels. The fungi include many common toadstool agarics and other basidiomycetes.^fn1 They tend to choose particular host trees: a whole genus (or subgenus), Alnicola, favours alder. They are particularly efficient on acid soils, where they recycle nitrogen and phosphorus out of fallen leaves. Even minerals in the pollen that birches, pines and oaks so copiously shed are not wasted but sent back into the tree by the root-fungi.⁶

The world’s best-studied mycorrhizal is probably matsutake (‘pine fungus’), Tricholoma matsutake, the fabulous edible toadstool which is part of any Japanese autumnal feast. (Slivers of it have passed down the author’s – alas unappreciative! – throat.) It grows on the roots of Japanese red pine (Pinus densiflora). Starting from a particular tree, it forms a slowly expanding shiro (a clonal mat of mycelium), which is poorly competitive against other fungi and microorganisms, and sensitive to the amount and composition of leaf litter. Fruiting is influenced by the age of the pines and the abundance of the evergreen oaks and hollies that form the understorey: there is a complex underground interaction between this and other mycorrhizals and litter-decomposers. Matsutake has been getting rarer, partly because the pineries have been destroyed by a nematode, but also because declining management has altered the balance between pines and shrubs. Since one toadstool can easily sell at the equivalent of £20 (retail in 2004), matsutake is more lucrative than timber.^fn2

Fungi, originating from trees, can transfer substance to other plants. This explains woodland plants that have no chlorophyll, such as bird’s-nest orchid and yellow bird’s-nest. Long regarded as saprophytes (feeders on rotten wood or leaves), they are really parasites on the ‘wood-wide web’ of ectomycorrhiza. Like the toadstools’, their substance was originally made by the trees. A mysterious speciality of Groton Wood (Suffolk) is Epipactis purpurata, a big orchid growing and flowering in the densest shade of lime where there are few or no other herbs. I never see it not flowering, and it never comes up twice in the same place; presumably it lives for many years without coming above ground. Some plants are pink instead of green, with little or no chlorophyll. I have long suspected that it is really a parasite on mycorrhiza.

Mycorrhizal fungi form a fourth component in the ecosystem, along with trees and shrubs, herbaceous plants, and bryophytes. They are a factor in the behaviour of woodland in relation to waterlogging. Anyone trying to start a new wood has to consider them, especially if the site is fertiliser-sodden farmland: cultivation and fertilising are disastrous for many mycorrhizas.

This is a record of my own long-term observations and those of visitors, including distinguished mycologists, at four sites in eastern England: two boulder-clay woods 2 miles (3 kilometres) apart but of different character, an ancient mixed coppice on a wide range of soils, and a series of plantations in a 200-year-old Breckland park (Table 4). (The development of fungal communities and guilds in Brandon Park is discussed on p.369f.)

Most of the classic sites for mycorrhizal fungi are on acid soils. Boulder-clay woods have a reputation for being poor in species. This is partly due to their fungi fruiting less often on calcareous soils; the difference is reduced by recording over many years. Mycorrhizal fungi are often associated with particular trees: Spooner & Roberts (2005) give many such associations. In a mixed wood it can be difficult to decide which tree goes with which fungus, especially as many mycorrhizal fungi are clonal: starting from one tree, they spread out into the root-space of different trees. In Table 5 I list the associations that I have been able to identify, which are summarised in Table 6. They confirm some of the traditional associations, but not all.

In Brandon Park, although the greatest number of mycorrhizals go with pine, beech and birch have disproportionately many in relation to the abundance of the trees. Birch and pine have many species in common; these are both pioneering and relatively arctic trees, but did not much occur together in the early Holocene of Britain. Ectotrophic mycorrhizas get about by spores; plantations of exotic trees tend to make do with the local ectomycorrhiza and later to acquire their own species. Thus the conifer plantations, in 200 years, have acquired five species of the conifer-specialist genus Suillus, including two specialists from the far-away homeland of larch.

In Bradfield Woods, with no conifers or beech, the predominant hosts are birch and hazel; alder has unexpectedly few specialists. There are surprisingly many species in common between Brandon and Bradfield: beech associates in Brandon (e.g. Laccaria amethystina) and even pine associates (e.g. L. laccata) occur with other trees in Bradfield.

In Hayley Wood, oak and hazel appear to be the predominant mycorrhizal hosts. In Buff Wood they are outdone by the small area of hornbeam, which has several specific mycorrhizals: some of these occur on beech in Brandon Park, even though hornbeam is more closely related to birch than beech.

Ash, maple, lime, hawthorn and blackthorn have few known associates. These have mainly vesicular-arbuscular mycorrhizas without visible fruit-bodies.

An anthropomorphic myth is that trees have a defined life span and die of ‘old age’. This may be true of some short-lived species. Most of the flowering cherries that were fashionable street trees of the 1930s are now dead; birch and aspen seldom reach a century. However, in a civilised country, trees are normally felled before they get far into middle age and become too big to be easily handled. The public rarely sees an old tree of a long-lived species.

Oaks are not immortal: they die at random from unknown causes. At Polstead Park in Suffolk a gigantic, spreading oak at the corner of the park, 17 feet (5.2 metres) in girth, suddenly died in 1991 (Fig. 20). Like all freestanding oaks with a big spread of branches it was young for its size, dating from c.1736. A row of pollard oaks round the edges of the park, nearly twice its age and very stag-headed, hardly altered at all. Life expectancy has little to do with age: if one must be anthropomorphic, the battlefield is a better analogy than the almshouse.

Why are trees not immortal? Every year trees have to lay down a new annual ring all over their trunk, branches, twigs and roots. Most trees reach their maximum leafage in late youth. Thereafter, taking good years with bad, the material available for making new wood is roughly constant, but it must be spread over an inexorably increasing area. Obviously this cannot go on for ever.

Trees can retrench and reduce the area to be covered: they shed redundant twigs and branches. Pollarding and coppicing are a major retrenchment, which resets the ageing process and prolongs the tree’s life. Most of the oldest oaks are pollards. Ash usually falls to pieces and disappears at around 200, unless growing in an adverse environment, but coppicing can extend this to at least 800. Some trees are self-coppicing or self-pollarding (p.64f). Lime, unusually, is able to miss one or more annual rings.

A veteran tree is one old enough to have gone through cycles of dieback and regrowth (Fig. 21). How old it is depends on the species: an 80-year-old aspen is a veteran aspen, whereas a 200-year-old oak is in early middle age.

Life expectancy depends more on size than age. Other things being equal, adversity, which slows a tree’s growth, will prolong its age. Very old trees occur on infertile soils or at high altitudes. The world’s longest-lived trees are the short, fantastic, gnarled bristle-cone pines at the limit of trees in the mountains of California. Among the oldest trees in Europe are the short, fantastic, gnarled cypresses at the limit of trees in the mountains of Crete (Fig. 22).⁷

Competition limits longevity by interfering with retrenchment. A tree that goes through a phase of dieback will not recover if a neighbour expands to fill the space. Although in other countries ancient trees can occur in woods, like the redwoods of California, in Britain they are normally freestanding.

Trees (God knows why) have evolved in favour of longevity. Decay is not a disease, but is part of that evolution and part of normal development. Any tree needs a mechanism to get rid of superfluous boughs formed when it was young. It must compensate for minor damage and survive injury by windblow (or fire, with fire-promoting trees).

Trees have no immune system and no wound-repair system. Vertebrate animals respond to injury by warring against the bacteria that invade wounds, and by regenerating damaged tissue. Trees react with a damage-limitation mechanism, walling off and bypassing both decayed and diseased tissue.

When a branch dies or is broken or cut, the tree lays down barriers in the surrounding wood which wall off a compartment, the size and shape of which are determined by the injury. Wood-rotting fungi (sometimes specific to particular living trees) invade the exposed surface and form a pocket of rot, which spreads up to the predetermined compartment boundary.

In Europe, research has been dominated by the interests of timber producers, who harvest trees before they are old enough to have incurred much injury. For them, decay is a nuisance like disease. Understanding of damage limitation comes from America, with its stronger tradition of arboriculture (caring for individual trees).

The barriers consist of pre-existing wood altered so as to block the spread of fungi. In broadleaved trees the water-conducting vessels are plugged with structures called tyloses: in conifers the microscopic bordered pits, acting as valves between the water-conducting cells, close. The wood becomes impregnated with tannins and other fungus-opposing chemicals. There are four types of barrier. The first three resist the spread of a fungus along the length of a stem or branch (important in blocking the spread of vascular diseases like Dutch Elm Disease), from one annual ring to another, or around the circumference of a stem. Type 4 is called forth by a major injury, and separates wood laid down before the injury from that produced after. This is the most effective type of barrier, often forming a bark-like lining to the interior of a hollow tree.⁸

Fungi themselves contribute to compartmentation. Some wood-rotters form a barrier of hard fungal tissue called a zone-plate on meeting the tree’s barrier. If they meet wood occupied by another individual fungus, each makes a zone-plate, with a narrow ‘demilitarised zone’ between (Fig. 23). Zone-plates in beech or hornbeam, formed by the fungus Ustulina deusta, appear in sections of the tree as single or double sinuous black lines. ‘Spalted beech’ (full of zone-plates) is treated as decorative by some furniture makers.

Most trees reach their maximum height and spread quite early in life: they pass through a period of being vulnerable to breakage by wind, but go on indefinitely getting thicker. When a big tree goes hollow it is getting rid of internal wood no longer needed to hold it up. Minerals in the decaying wood may be recycled into the tree’s roots. Although in Britain decay is a mainly fungal process, in lower latitudes termites help. In the Cape York Peninsula of tropical Australia most trees, even a few inches in diameter, are hollow (and are used by Aborigines for the musical horn called a didgeridoo).

American, European and Australian trees – broadleaves, conifers and eucalypts – have this mechanism. In the great storm of 1987, most of the trees that broke were big, young and perfectly sound. Many a rotten or hollow tree remained upstanding surrounded by younger trees that were broken or uprooted.

Trees exhibit compensatory growth, laying down extra wood where needed at weak points (see Fig. 26). The base of a big tree tends to be a weak point, where butt-rotting fungi can spread up from root injuries, and where the tree is unlikely to survive breakage unless it is a lime. (I have known a big oak break at the base, the wood having been turned by a white-rot fungus into a substance still identifiable as oak, but with the consistency of Camembert cheese.) Buttress roots may compensate for butt-rot: many a great oak is held up for centuries by its root buttresses, the centre of the base having disappeared.

Most decay fungi are symbiotic rather than pathogenic, but some can be either. Silver-leaf fungus, Stereum purpureum, is a weak wood-rotter, especially on birch, but on fruit trees it can enter through pruning wounds and secrete chemicals that poison the tree. All the almond trees around where I lived in Norwich died of it during ten years or so. After World War I the Government, thinking it could vanquish the fungi, made silver-leaf illegal. Forestalling silver-leaf is probably why arboriculturalists used to cover pruning wounds with tarry preparations ‘to prevent decay’. In reality it is impossible to prevent decay, but there is something to be said for encouraging decay by symbiotic rather than pathogenic fungi.^fn3

How effectively compartments confine pathogens depends on the fungus and the tree. The bracket-fungus Inonotus hispidus, a heartwood-rotter of ash, gets through compartment walls by growing in winter when the tree’s reactions are slow. Thus a heart-rotted ash can withstand one winter storm but break in a less severe storm the following winter. Ustulina deusta is well contained by beech, but can penetrate the compartment barriers of lime.⁹

A pollard tree renews its Type 4 barriers each time it is cut. Inside a hollow tree there may be the shells of one or more previous barriers, superseded by the present lining of the hollow; they weather out as the surrounding, less rot-resistant wood decays. Roots can spread downwards into the rotten wood and debris that fill the cavity, extracting minerals. (Compartment barriers, formed when the tree was alive, occasionally weather out from decayed timbers in ancient buildings.)

Much of the value of trees as a habitat for other creatures attaches not to all trees, but to those few that are old enough to have hollow interiors. Beneficiaries include:

hole-nesting birds;
bats, using cavities and cracks of different sizes for roosting, hibernation and breeding;
beetles, flies and other insects of rotten wood, especially of red-rotted oak, willow or hawthorn altered by Fistulina hepatica;¹⁰
insects associated as symbionts or predators with bracket-fungi;
invertebrates living on damp debris accumulated inside pollards;
flies that breed in wet rot-holes, or on ‘slime tracks’ of exudate (especially those produced by bacterial wetwood),¹¹ dribbling down the outside of a tree;
insects under old dry bark;
spiders associated with these, and insects that feed on the spiders’ leavings;
specific lichens on old dry bark under overhangs;
bryophytes and other lichens on rain-tracks or weathered-out compartment boundaries.¹²

A single 400-year-old oak, especially a pollard with its labyrinthine compartment boundaries, can generate a whole ecosystem of such creatures, for which ten thousand 200-year-old oaks are no use at all.

Most native trees produce annual rings, whose structure is diagnostic of the species (p.242ff). The age and growth rate of a tree are given by counting the rings on a stump or in a core taken with a hollow borer. (Many trees, especially in the tropics, have no annual rings; one should be cautious with trees like lime that can miss rings, or with foreign trees that produce extra ‘false rings’.)

Oak, by far the commonest timber in historic buildings, has well-defined rings that vary in width from year to year. The variation depends partly on weather; good and bad years are repeated in most trees within the same geographical region. Posterity will recognise the late twentieth century by the sequence of a bad year (1975), a worse year (1976), then 12 unremarkable years, then two bad years (1989–90), three unremarkable years, a bad year, an average year (1995), an average year, a rather bad year (1997), and so on. By comparing the early years of a living tree with the later years of a timber from a historic building, and by comparing its earlier years with the later years of an older timber or a log buried in a bog, these patterns have been extended back, even to the Bronze Age.

This procedure has grown into a considerable science. It is used to date historic buildings: the journal Vernacular Architecture publishes results of such work. The determined amateur can try it: it helps to have a travelling microscope to measure the rings to within 0.1 millimetre. There are computer programs to remove the effects of long-term trends, as the tree grows older or as neighbours compete with it, leaving the year-to-year variations that are useful in dating.

Tree rings have other uses. Because weather varies from region to region, the provenance of a timber can sometimes be determined: if the sequence matches a master curve from Poland rather than England, this is evidence that the sample is of Baltic oak. By removing year-to-year variation, leaving the long-term trends, it has been possible in America to use growth rates as a measure of climate change.

To get a result one normally measures at least 100 rings, preferably from each of several contemporary trees or timbers. Tree rings are affected by other factors besides weather, such as defoliating caterpillars.^fn4 In view of the statistical ‘noise’ introduced by unknown factors, it is surprising that the method has been so successful and so seldom at odds with dating by other means.

A now familiar sight is a small tree trunk with horizontal rows of holes round the circumference, sometimes repeated from top to bottom. Americans recognise these as the work of sapsuckers (Sphyrapicus), a genus of woodpecker-like birds that peck holes in trees and then come back and lick up the exuding sap. Sapsuckers, however, are absent from Europe.

The trees most often affected are lime, elm and oak, typically stems about 4 inches (10 centimetres) in diameter with nearly smooth bark (Fig. 24). The holes are soon overgrown by new annual rings, but the scars persist for many years in the bark and permanently as little irregularities in the wood (to be treated as decorative by future veneer-cutters?).

This minor damage seems to be done by the great spotted woodpecker; other woodpeckers may be involved on the Continent. It is curious that such a conspicuous activity was not noticed in bird or forestry books down the centuries. I began to see it around 1970, but there are scattered records (more numerous in Central Europe) back to the 1930s. Have woodpeckers only recently taken to sapsucking?

As Edlin noted, there is little hope of studying fire and trees in England. Readers visiting the Mediterranean, North America or the Scottish Highlands may encounter it, and in Australia certainly will.

When fires happen depends on weather and on sources of ignition; but whether vegetation will burn at all depends on the plants. Trees are combustible by adaptation, not misfortune. Pines and eucalypts make flammable resins and oils, and it is their business to burn from time to time and destroy their less fire-adapted competitors. More subtle adaptations can promote fire. Fallen leaves may lie as a loose, airy litter; shed twigs may lie for many years without rotting; the trees may cast a light shade that allows flammable shrubs to grow underneath. Fires are reproducible: it is unusual for vegetation to burn that has never burnt before. Combustibility apparently evolved with some of the earliest trees in the Devonian; so presumably did adaptations to take advantage of it. Charcoal from Carboniferous forest fires is a major constituent of coal. Since the late Tertiary, fire has come to be a factor second only to climate in determining the world’s ecosystems.¹³

Lightning often strikes a dead tree and sets fire to loose bark or dry debris in its hollow interior. Even now, in a tidied-up landscape, most fires in the Rocky Mountains are attributed to lightning. Climate can encourage fire, but does not cause fires unless the vegetation is fire-promoting.

Fires vary in intensity, duration and effects. A fire in ground vegetation or leaf litter is usually relatively cool. Its effects tend to be long term, killing seedlings and young trees of fire-sensitive species like beeches and maples. A crown fire, which happens with fire-promoting trees, is much grander: flames climb into the treetops and leap from tree to tree.

Certain plant families – pines, eucalypts, heathers, grasses and palms (Fig. 25) – are fire-adapted. Fires seldom happen without them. Some species are easily killed, but their seeds are stimulated to germinate by heat, and the tree starts producing new seeds when young: Aleppo pine in the Mediterranean can regenerate if fires are only 15 years apart. Others are killed to the ground, but sprout from the base as in coppicing, for example strawberry-tree. Others have thick bark or heat-resistant cambium, and a full-grown tree is little affected: the supreme example is cork oak. Some pines have cones that will not release their seeds unless heated. Many Australian plants have seeds that will not germinate unless exposed to smoke or water that has passed over charred wood: visitors to Perth Botanical Gardens are sold packets of seed and bottles of smoke-water to persuade them to germinate.

In Europe the pines of the north and the Mediterranean vegetation of the south are fire-adapted; not so the deciduous forests of the middle. Similar-looking ecosystems in North America are more fiery, with extensive pines and tall Ericaceæ and Rosaceæ, big flammable shrubs that accumulate fuel. America, however, also has anti-fire ecosystems. Sugar-maple, a strongly dominant tree whose shade-resistant seedlings are killed by even a weak fire, has dead leaves that compact into a dense, incombustible, ground-hugging mat. In turn the monocotyledonous spring flowers have upright, sharp-pointed leaves that pierce this papier-mâché-like layer. Fire may be linked to gregariousness: adaptations both to fire and to lack of fire work only if there is a considerable area of vegetation with the same characteristics.

Australia is the Planet of Fire. Except in the small area of rainforest, fire is as necessary to Australian native vegetation as rain to Britain. All the thousand species of Eucalyptus that dominate Australian vegetation appear to be fire-adapted. One cannot live long in Sydney without witnessing the awesome spectacle of a eucalyptus crown fire, or the nonchalant way in which the trees carry on growing afterwards. These eucalypts, although they may not have obviously thick or fireproof bark, survive and grow new twigs to replace the burnt ones. Some shed their outer bark to fuel a second fire a year later, mopping up any competitors that the first fire missed. Other species are killed to the ground but sprout, usually from a lignotuberous stool that gets bigger at each fire cycle. Western Australia has many bizarre sights: none more so than an old marri (Eucalyptus calophylla) converted by repeated fires into the likeness of a pollard, a pillar of charcoal sprouting leaves. Possibly the grandest fire anywhere on the planet in human history occurred in 1939 in some of the world’s grandest trees (E. regnans and E. delegatensis) between Melbourne and Canberra. These 300-foot (90-metre) eucalypts – themselves arising after a fire centuries ago – were killed, but their tiny seeds germinated to form a new generation, by now over half the height of their parents. All these notoriously do not regenerate without fire; apparently they need fire to set back antagonistic fungi in the soil.¹⁴

Fires can occur in ordinarily fireproof vegetation if a disturbance causes fuel to be drier or more compacted than usual. Debris left after logging is famously prone to ignition. In Borneo, fires occurred in tropical rainforest after drought killed some of the trees and a second drought dried them; this is probably a recurrent though rare event, and most living trees more than 28 inches (70 centimetres) in diameter survived the fire.¹⁵

The visible effects vary. Charred stumps and logs retain a coating of charcoal for several decades, and when it weathers away it leaves a smooth surface with a distinctive texture. On Mount Athos, the sacred mountain in north Greece, it was easy in 2001 to recognise traces of the great fire of 1991, but there were fainter traces of two previous fires beyond the present monks’ memory.¹⁶ Citizens of Sydney learn to recognise the stages whereby trees regain leafage burnt in a crown fire, as new growth sprouts from the bigger, living boughs and twigs.

A symptom of fire is a charred cavity at the base of a tree, usually on the uphill side, which is enlarged by successive fires. In Tasmania most of the big savanna eucalypts have conical bases (depicted by early English artists) resulting from compensatory growth around cavities eaten out by grassland fires in the nineteenth century (Fig. 26).

Fire, like coppicing, stimulates a profusion of herbaceous plants, many of them from buried seed laid down after the last fire. This can be seen after fires in English gorse or Mediterranean maquis. Often the part of a forest richest in plant life is the part that has recently been burnt.

Forest fires favour deer, which can increase tenfold after a big fire. Human burning practices are the oldest detectable form of land management, and still almost the most widespread. The object is not to get rid of forest. It may be to make better pasture for wild or tame animals, or to create areas of improved pasture on which to find the beasts when needed, or to help in catching beasts by persuading them to leap over cliffs. There is abundant ethnographic evidence, especially from North America.¹⁷

Europeans (unlike Native Americans and Australian Aborigines) fear fire, and when they colonise fiery regions try to suppress it. A by-product of the fire-fighting industry is that small, controllable fires are replaced by vast, uncontrollable conflagrations stoked by many years’ accumulated fuel. However, if fires are delayed beyond the usual interval, the site may be invaded by fire-sensitive trees that alter the chemistry and structure of the vegetation so that it never burns again.¹⁸ The sugar-maples of the southeastern United States have been suppressing shortleaved pines and the beautiful, fire-dependent ecosystems that go with them; the rainforests of eastern Australia have been expanding into the adjacent eucalypts.

I remember when ‘forest fires’ were much feared in Britain, and in retrospect can see why. In the Forestry Commission’s youth, young conifer plantations, packed with bracken and dry grass, were an admirable fuel; when he-men smoked and trains shovelled burning coals on to the track there was no lack of ignition. A ploughed strip was maintained where railways adjoined plantations. Even thunderstorms set plantations on fire, and the rain often did not extinguish them.¹⁹ These conditions have passed away.

English native woods burn like wet asbestos. Even in the great summers of 1975 and 1976, when extreme drought coincided with a fashion for burning stubble fields, I never heard of a grown-up native wood catching fire – nor in the summers of 1995 and 2003. It happens that Nature has given England very few fire-promoting plants – heather, gorse, dead bracken, dry reed-beds, dry grasses – none of which is a tree. Fires happen in certain types of non-woodland vegetation, but run out of fuel and go out on reaching a wood (Fig. 27).

Although English woods appear to be immune from fire, trees may be involved in a fire in surrounding bracken or heather. Bracken in shade makes too little substance to sustain a fire, but a wood can burn a few years after it has been felled and the bracken thereby stimulated. In the exceptional summer of 1975–6 I encountered four instances of bracken or bramble burning in a sparse or recently felled wood: I doubt whether they did permanent damage. In 1326–7 there was a sale of £1. 2s. 9¼d.-worth of burnt underwood in a wooded part of the Forest of Dean, resulting from ‘an unfortunate fire from a certain charcoal pit setting light to the bracken’; the wood was expected to grow again.²⁰

In sparse oakwoods with plenty of bracken it is worth looking for charred cavities on the uphill side of tree bases. Professor Paul Mellars tells me the strange story of Treeton Wood near Sheffield, which he remembers as having bluebells and bracken, but which by 2000 was reduced to stumps as a result of a fire. This may be a repeated event; another of the three ancient woods in Treeton was named ‘Burnt Wood’ before 1860. All three, according to the Ordnance Survey, had conifers planted in them, which may have predisposed to fire.

Lack of fire as a significant factor in English woodland is not a legacy of past management resulting in an unnaturally tidy wood lacking fallen branches. It is due to the anti-fire adaptations of native trees. Most have a leaf litter that compacts into a soggy mass. Fallen branches quickly rot: combustible deadwood never accumulates to provide fuel. This leaves only trunks, especially oak, that are too big to burn on their own.

Burning is a method of rejuvenating (but not creating) heathland. In wood-pasture, grasses, heather and bracken provide fuel, and trees are involved only incidentally. Ancient trees are easily killed by fire, especially if hollow, which is a significant conservation problem: many of the ancient oaks of Ashtead Common (Surrey) have perished in a succession of fires. (Fires do not solve the problem of conserving heathland; birches spring up on the burnt surface and exert their anti-fire properties, shading out the heather to produce an undistinguished secondary wood.)

So much for England. In Scotland the native pine, like all the world’s pines, is fire-adapted (Chapter 16). The ‘natural’ frequency of fires in the pinewoods of Scandinavia is said to be about every 80 years. Although fires in Scotland were probably less frequent, they are known from most Caledonian pinewoods, in some of which (as in Glen Tanar) areas of even-aged pine result from historic fires. Young pines are killed by burning the heather around them (as well as by too much grazing); occupational burning of moorland is probably one factor holding back the pinewoods from spreading.

In Ireland there is the strawberry-tree, now almost confined to the Killarney area, but which from place-names and other evidence was more widespread in historic times. This small, light-demanding tree is an outlier from the Mediterranean, where it is often dominant in maquis. It is flammable and fire-adapted: it makes Corsica such a fiery island. In Killarney it grows on cliffs and at the edges where oakwoods meet bogs, which have a history of burning. Its decline cannot be, as often said, due to people cutting it down, for it sprouts vigorously after felling as well as burning. Does lack of fire play a part?

Too many vegetation historians appeal airily to ‘forest fires’ without giving actual examples (localities, species and dates) of fire occurring in the kind of forest being considered. It is no good citing analogies from American Indian practices: North America is a different continent, has different trees, and is more flammable than this part of Europe can ever be. Grass fires in savanna are not evidence that forests will burn. One still encounters the nonsense that prehistoric people ‘cleared the forest by fire’ to create fields: nonsense because even in the most favourable circumstances fires may kill trees but do not destroy them, and there remains the immense task of getting rid of the dead trees and roots.

Charcoal occurs in prehistoric peat deposits: conscientious pollen analysts record the varying amounts. The fragments are often too small to identify, being charcoal dust blown in from a distance. In countries with flammable trees these may have come from forest fires. In England the presence of charcoal (unless it came from a very local fire on the peat surface itself) probably means that what burnt was not forest but heath or savanna, with grass, heather or bracken. Pollen deposits, moreover, are usually from wetlands, which can be more combustible than the rest of the landscape; in reed-beds, especially, the standing dead stalks are often burnt in spring by people for reasons of land management.^fn5 The relation between charcoal and a part-savanna wildwood, on Vera’s model, needs to be explored.

fn1 Two other types of mycorrhiza are associated with Ericaceæ and terrestrial orchids.

fn2 I am grateful to Professor Hideo Tabata and colleagues for an initiation into pityomycology.

fn3 In America, wounds on live oaks are painted to deter insects that carry oak wilt.

fn4 I am told that at conferences in this field high words, and occasionally blows, used to be exchanged between the supporters of weather and of caterpillars.

fn5 I am grateful to Professor Paul Mellars for discussing this point concerning the Mesolithic site at Star Carr, Yorkshire.