Taxus baccata is a native evergreen non-resinous gymnosperm tree up to 20(−28) m, often with multiple trunks and spreading, rounded or pyramidal canopy. Capable of producing leafy branches from old branches and trunks, and sometimes from stools. Root system shallow with extensive horizontal roots, often above ground on calcareous substrates. Bark reddish-brown, thin, scaly. Branches long, not whorled. Twigs green and irregularly alternate; buds very small, bud scales dark-brown, rounded, imbricate and closely appressed. Leaves, spirally attached but on lateral shoots twisted more or less into two ranks, can live for up to 8 years; 1–3(−4.5) cm long and 2–3 mm wide, parallel-sided, shortly stalked, mucronate, dark glossy green above, paler and yellowish beneath with two pale stomatiferous stripes; midrib prominent on both sides, margins recurved, tapering to a petiole-like base. Normally dioecious, rarely monoecious; reproductive structures green, borne in leaf axils near the end of the previous summer's growth. Male parts in small cones (simple strobili; Mundry & Mundry 2001) 2–3 mm diameter, made up of 6–14 stalked peltate microsporophylls each with 4–9 pollen-sacs; strobili in clusters of up to 20–30 near branch ends. Female structures (ovuliferous brachyblasts sensuStützel & Röwekamp 1999) 1.5–2.0 mm long, solitary or in pairs in leaf axils on underside of shoots, not forming cones, consisting of several imbricate scales, the uppermost fertile, bearing a single ovule. Seed ovoid, smooth and shiny, brown-yellow, 6–7 × 5 mm at maturity, with a tough seed coat, partly surrounded by a fleshy red aril typically 9 × 7 mm which falls with the seed at maturity, the ‘fruit’ ripening in the first year.
Taxus is a genus of around seven closely related species scattered through the northern temperate region. There is some indication from morphological and chemical characteristics that the different species are sufficiently closely related that they constitute geographical variations of T. baccata (Hess et al. 1967 cited in Voliotis 1986; Burns & Honkala 1990; Dempsey & Hook 2000). Indeed, the heartwood constituents of T. baccata, T. brevifolia, T. cuspidata and T. floridana have been found to be almost indistinguishable chemically (Erdtman & Tsuno 1969). Most species contain the alkaloid taxol and although Dempsey & Hook (2000) found wide variation in morphological characteristics and chemical features (paclitaxel content) between varieties of T. baccata they found no sufficiently distinctive interspecific differences of taxonomic value between different species. They thus question the correct identification of plant material used in many pharmacological studies.
Taxus baccata has a stress-tolerant life strategy sensu Grime (1979), being slow growing, slow to reach maturity (c. 70 years), long-lived (> 1000 years), shade-tolerant but can withstand full sun; it can grow on a wide range of soils and produces strong decay-resistant wood (Lilpop 1931; Brzeziecki & Kienast 1994). It forms dense stands in oceanic climates but increasingly becomes an isolated understorey tree in more continental climates.
This evergreen poisonous tree is local in Britain but may form nearly pure woodlands on chalk and limestone in England.
I. Geographical and altitudinal distribution
In Britain, Taxus baccata woodlands are almost wholly confined to the chalk of the South Downs and, to a lesser extent, the North Downs in south-east England especially along the sides and bottom of dry valleys and on scarp slopes (Tittensor 1980; Rodwell 1991). There are also examples on the Magnesian Limestone of County Durham in the north-east, the Carboniferous Limestone around Morecambe Bay in the west and on the limestones of south-west Ireland, especially the Killarney woodlands (Kelly 1981). Further small, natural-seeming stands of yew are found in Ireland on limestone outcrops at Lough Derg, north Co. Tipperary, Cahir Park, south Co. Tipperary and Garryland Wood, south-east Co. Galway (McEvoy 1943).
Individual yew trees are more widespread, occurring naturally in woods and on cliffs, including sea cliffs, especially on neutral to alkaline soils. They are notable on the limestone of the Peak District at the southern end of the Pennines, at Borrowdale, on Conzie Scar near Kendal (Elwes & Henry 1906) and widely distributed in the west and north of Ireland (Webb 1977). Apparently native T. baccata individuals are also known from Perth and Argyll (56° N) on to the Inner Hebrides and Orkney although whether yew is native to Scotland has been debated (Dickson 1994). The natural distribution is clouded by the extensive planting of yew trees, especially in churchyards, and it is virtually impossible to separate native and alien records; however, the occurrences mapped in Fig. 1 give the best information available.
Taxus baccata extends northwards to c. 63° N in Norway and Sweden (Vidaković 1991), eastwards to Estonia, Poland, to the Caspian Sea and Turkey, and southwards to Greece (including the islands of Euboea, Thasos and Samothraki; Voliotis 1986), northern Spain (Peñalba 1994), Portugal and into Algeria (Vidaković 1991). Large populations are found in the Ukraine, Poland, Hungary, Slovakia, Romania and the Caucasus Mountains (Bugala 1978). It is absent from the most continental climatic regions of Europe and also northern, south-eastern divisions of Russia, Crete, Faeroes and Iceland, and is almost extinct in the Azores. The distributions of Taxus species, including T. baccata, are given in Fig. 2.
In the British Isles, T. baccata grows from sea level to 425 m in England and Wales and to 470 m on Purple Mountain, Kerry (Alt. range Br. Pl.). In the north of its range, including Britain, T. baccata grows only at fairly low altitudes whereas in the south it is primarily a montane tree growing only on higher mountains in the Mediterranean region, although in the Caucasus Mountains it grows from sea-level to 2050 m. Maximum altitudes: south Slovakia, 660–1000 m; Alps, 1100–1400 m; Iran, 1400 m; Turkey, 1400–1900 m; Pyrenees, 1400–1650 m; southern Spain, 1600–1900 m; Carpathians, 1660 m; Sardinia, 1700 m; Macedonia, 1800 m; central Greece, 1950 m; northern Africa, 2000–2500 m, and Caucasus Mountains, 2050 m (Elwes & Henry 1906; Bugala 1978; Król 1978; Melzack 1979; Voliotis 1986; Hartzell 1991; Vidaković 1991; Schweingruber 1993).
(a) climatic and topographical limitations
Taxus baccata grows best in the high humidity of mild oceanic climates. In England it thrives particularly in areas with relatively mild winters, abundant rainfall, high humidity and where mist frequency is greater than average, as in the coastlands of Hampshire and Sussex and lowland parts of the Lake District (Tittensor 1980; Melzack & Watts 1982b). Conversely, severe winter cold or strong cold drying winds in the spring appear to restrict yew growth on exposed sites in the UK (Voliotis 1986).
Taxus baccata sensu stricto is classified as European Temperate by Preston & Hill (1997). In mainland Europe, the ecological barriers limiting distribution are low temperatures in the north, severe continental climate of the Baltic countries and Poland in the east, long droughts in Turkey to the south-east, and drought and high temperatures in north Africa (Kamppa 1926; Król 1978; Środoń 1978). Near these climatic extremes, T. baccata becomes more restricted to moist niches, such as near marshes and bogs on the Baltic coast (Król 1978), rock crevices, or as an understorey tree gaining protection from low humidity and frosts (Sokolowski 1921; Hoffmann 1958; Faegri 1960; Tittensor 1980; Peñalba 1994; Saniga 2000). In mountainous areas of Europe, yew tends to grow on the shaded north-western or north-eastern slopes, where under deciduous or mixed stands there is an oceanic-like climate (Markiewicz 1978; Schweingruber 1993).
Taxus baccata woods on the South Downs occur on slopes of almost any aspect but primarily linked to south, east and, to a lesser extent, west slopes where there is high insolation and wind-exposure (Watt 1926; Williamson 1978). Slope is rarely limiting; stands of yew occur on moderate to very steep slopes (Rodwell 1991) and, further north, cliffs are a common habitat (Pigott 1969). On the western South Downs, 80% of yew woodlands were investigated by Tittensor (1980) and found to be between 90 and 250 m altitude, 62% on steep slopes, 57% along the main ridge and 90% on chalk. Moreover, 80% of the woods occur in the area of maximum rainfall (> 1000 mm year−1).
Stands of yew are typically associated with limestone slopes carrying shallow dry rendzinas (Rodwell 1991). Soils are normally grey rendzinas of the Upton series, thin, often rich in downwashed flints and poor in earthworms. Yew grows equally well on thin, warm chalk soils, limestone pavement and fen peat (Williamson 1978). In south-west Ireland, yew grows on Carboniferous Limestone pavement with a soil pH in the hollows of 6.8 ± 0.11 (SE) (Kelly 1981). Along valley bottoms, T. baccata woodland may extend on to somewhat deeper and moister soils.
Although the majority of natural stands in Europe grow on calcareous soils, yew will also grow on almost any soil, including siliceous soils derived from igneous and sedimentary rocks (Elwes & Henry 1906; Watt 1926; Webb 1977; Król 1978; Kelly 1981; Vidaković 1991). Thus in south-east England, yew grows on the chalk of the North and the South Downs, the Weald Clay and the sandstones of the Lower Greensand and the Central Weald. In the west and north of Ireland, individual trees occur on rocky terrain, siliceous as well as limestone, and in Killarney woods, individuals are found on Devonian sandstone (Kelly 1981). Voliotis (1986) states that T. baccata is favoured by slightly calcareous soil in northern Europe and siliceous soils in southern Europe. Watt (1924, 1934a,b) noted that yew is a successful pioneer on highly calcareous, exposed sites of the South Downs, but on more acidic plateau soils T. baccata is a scarce but local feature with low constancy and frequency values. Across Europe, best yew growth (in terms of rates of growth and largest dimensions) is seen on deep, moist sandy loams and well-drained clays, and worst on dry, rocky and sandy soils where it occurs as scattered and stunted individuals (Webb 1977; Bugala 1978).
Yew generally is absent from wet soils, such as wet acidic peat and wet clay (Król 1978; Williamson 1978; Ellenberg 1988), although it does grow well upon calcareous fen peat, and indeed is a significant species in the wood peat accumulations of the East Anglian fens (Godw. Hist.). Even here, however, it is susceptible to poor drainage.
Substratum conditions are known to be altered by T. baccata. Howard et al. (1998) compared soil physico-chemical conditions under oak (Quercus) and yew trees growing on the same soils. Under yew, the humic acids were more oxidized; ash content was less under yew (55.6%) than under oak (73.8%); a greater total C, N and Ca was found under yew (23.7%, 1.3% and 2.1%, respectively) than under oak (13.1%, 0.8% and 1.1%), attributed to the absence of large earthworms under yew. Soil pH was not, however, significantly affected by tree species.
In Britain, T. baccata forms a remarkably species-poor woodland with yew as the only constant species, typically forming a closed, very dense canopy (W13 in the National Vegetation Classification; Rodwell 1991). The woodland forms ‘sometimes quite extensive stretches of striking floristic poverty and uniformity’ especially on the warm and sunny south-facing slopes over shallow limestone soils (Rodwell 1991). Yew thrives better than any other tree on steep calcareous slopes and so few other trees occur; when they do it is usually as scattered individuals, notably Sorbus aria, but also Fraxinus excelsior and widely scattered individuals of Fagus sylvatica, Acer pseudoplatanus or Quercus robur. Shrubs are rare except in gaps and edges of stands, seldom form a true shrub layer, and are usually restricted to poor specimens of Sambucus nigra and very occasional Ilex aquifolium or Crataegus monogyna. Buxus sempervirens is a rare associate. The field layer is, as would be expected, extremely sparse usually consisting of patchy herbs with extensive bare ground. The most frequently encountered species are Mercurialis perennis with very occasional Arum maculatum, Brachypodium sylvaticum, Fragaria vesca, Glechoma hederacea, Hedera helix, Rubus fruticosus agg., Urtica dioica and Viola spp. (Rodwell 1991). This is caused not just by shade but by the intense root competition in soils which are often already dry. Bryophytes are equally poorly developed in the UK, in striking contrast to the Killarney yew woods (see below).
Rodwell (1991) recognizes two subcommunities. The Sorbus aria subcommunity (W13a) contains this tree as a fairly frequent associate in the canopy, at about 2–3 trees ha−1, and even more rarely, Fraxinus excelsior, Fagus sylvatica, Quercus robur or Acer pseudoplatanus. In Co. Durham, Hulme (1996) also found Corylus avellana and Ulmus glabra in all four yew woodlands visited. Sambucus nigra is the usual shrub although in a few stands in the south-east, notably Box Hill, Surrey, Buxus sempervirens can be locally important and can grow up to form a dense shrub layer or even part of the canopy. A field layer is usually absent or at most very sparse – scattered bryophytes, bare soil and litter. In the Mercurialis perennis subcommunity (W13b), yew dominates with very infrequent Sorbus aria and Fraxinus. The canopy tends to be more open, allowing a more frequent and somewhat denser growth of Sambucus beneath, sometimes with sparse Ligustrum vulgare, Euonymus europaea, Cornus sanguinea and Clematis vitalba or Tamus communis. The field layer is a little less sparse than in the Sorbus aria subcommunity (although most of the ground is still bare) with Mercurialis perennis as a patchily abundant constant component of the field layer, sometimes with Atropa belladonna, Brachypodium sylvaticum, Fragaria vesca, Inula conyza, Iris foetidissima, Rubus spp., Urtica dioica and Viola spp.
Yew saplings are also a component of the Crataegus monogyna–Hedera helix scrub (W21). Young yew trees are found in the Brachypodium sylvaticum subcommunity but is a more major component of the Viburnum lantana subcommunity confined to free-draining calcareous soils, and best represented on the chalk of the south-east. Taxus baccata and Sorbus aria are the most frequent trees. Yew is associated with C. monogyna on deeper, moister soils on gentle, sheltered slopes but especially with Juniperus communis bushes on shallower, drier soils on steeper and more exposed sites where yew has a high constancy (Watt 1934a; Rodwell 1991). This reflects the need of yew for nurse plants in establishing new trees and leads to seral development.
The juniper sere starts with juniper establishing (in windier areas than Crataegus) which is invaded by yew, leading to a yew scrub with associated Cornus sanguinea, Prunus spinosa, Rubus fruticosus, C. monogyna, Ilex aquifolium and Sambucus nigra (Watt 1924, 1926, 1934a). Owing to its extreme tolerance of shade and the dense shade it creates, yew eventually outcompetes other species and develops progressively into pure yew woodland, often with the dead woody remains of the former juniper scrub beneath. There can be occasional I. aquifolium and rare straggling C. monogyna and Prunus spinosa, with an occasional Sorbus aria above the yew canopy, as at the head of Kingley Vale, Sussex. The Sorbus initially survives by outgrowing the yew in earlier years to persist as emergents or invaders of gaps. But as they die they are not replaced and the canopy of yew fuses into a virtually continuous layer (Watt 1926). The association of juniper and yew is not solely found in the south-east; Larson et al. (2000) records the two species growing together on the sea cliffs of the Great Orme and the Little Orme of northern Wales in 1997.
The hawthorn sere is similar, although C. monogyna is less good at providing protection for yew seedlings from grazers, even when mixed with Prunus spinosa. Moreover, on the deeper, moister soils where Juniperus is replaced, other trees, notably Fraxinus excelsior, increase in importance as invaders, leading to a more species-rich scrub. Hulme (1996) shows data which suggest a Crataegus/Fraxinus successional sequence in Co. Durham yew woods, with yew eventually killing out the scrub elements.
Although yew is capable of invading grassland (Williamson 1978), either juniper or hawthorn appears essential for its regeneration (Watt 1926, 1934a; Williamson 1978). This association between yew and shrubs has been interpreted as a requirement for establishing yew to have shade and/or protection from browsing vertebrate herbivores (Watt 1926; Williamson 1978). As such, juniper may be the more favoured nurse plant because it grows on soils more suitable for the domination of yew but also because juniper provides a better defence against herbivores. Hulme (1997) and García et al. (2000) point out that scrub will attract seed-dispersing birds and so increase yew seed deposition while, conversely, attracting high seed predation from rodents under the shelter of scrub. Nevertheless, dispersal appears to direct seeds to microhabitats most suitable for seedling survival, identified by García et al. (2000) as a favourably moist and nutrient-rich microclimate (especially in the face of summer drought), protected from large herbivores.
Tittensor (1980) working on the South Downs found that twice as many Recent as Ancient Woodlands contained yew, and that yew regeneration was far more common in Recent woods. Those Ancient Woodlands containing yew had few canopy yew and none with yew as a dominant species. Combining this with the estimated young age of woodland yews, she and others (e.g. Watt 1926; Williamson 1978) point out that the South Downs yew woods largely result from the abandonment of land during the last two centuries as a result of the Napoleonic wars, agricultural neglect in the 1920s and myxomatosis. Similar expansion in Irish yew woodlands has been seen following a cessation in grazing or human disturbance (Watts 1984; Mitchell 1988; O’Connell et al. 1988) although some, such as the Reenadinna Wood at Killarney, are estimated to have developed 3000–5000 years ago (Mitchell 1990b). Tittensor (1980) also suggests that because many of the woodlands straddle parish boundaries, old marker trees (which may be centuries or even millennia old) may have been the seed parents and the cause of yew woodland development which is not seen in other areas of abandonment.
Yew is extremely shade tolerant; nevertheless regeneration within mature yew woodland is rare in Britain and across Europe (Król 1978). The replacement of individual trees depends entirely on the development of a shrubby sere in gaps left by dead yews but primarily around the edge of existing woodland (Watt 1926; Williamson 1978). Regeneration to adult size may take as long as 75 years and the maintenance of yew woodlands is possible only owing to extreme longevity of individual trees and yew's tolerance of shade (Watt 1926; Tittensor 1980). Newbould (1960) concludes that yew woods are single generation stands ‘moving’ across the landscape by edge regeneration. If shrubs fail to develop densely enough, rabbits may prevent regeneration and the woodland degenerates by a failure to recolonize gaps formed by the death of old yews, leaving individual relict trees until eventually they die leading to ‘abandonment of the area’ (Watt 1926). Scrub can reform and the cycle would begin again with slow yew recolonization. Newbould (1960) suggests that yew woods may be considered as a type of secondary seral scrub that will progress to a ‘woodland community as yet unknown’ once the life-span of its component yews is reached.
Individual trees are also found much further north as a natural part of stands of Fraxinus excelsior–Acer campestre–Mercurialis perennis (W8) woodland on Carboniferous Limestone in Wales, Derbyshire, the Pennines and around Morecambe Bay (Rodwell 1991). Here the yew may form part of the canopy, a distinct lower tier to the canopy as in Matlock Dale in Derbyshire, or exist as scattered saplings or isolated specimens on cliffs as in Dove Dale Wood, Derbyshire (Pigott 1969). Although Rodwell (1991) considers these northern examples as part of the mixed deciduous woodland of W8, Hall et al. (2001) report that it is now common practice that they be included within the yew woodlands proper (W13).
Yew establishment is hindered and saplings often die or show poor growth with few leaves when beneath the shade of beech. In this way, beech may prevent the succession to yew. Yew does better on drier, more exposed conditions and this may explain why yew dominates here but beech is the dominant on the deeper soils of Fagus–Mercurialis woodland (W12). The circumstances under which yew develops from being an occasional component of woodland to becoming the dominant species remain poorly understood.
Yew is an occasional and sometimes prominent associate in various kinds of deciduous woodlands, most notably those in the Fagion alliance, especially the three kinds of beech woodland, Fagus sylvatica–Mercurialis perennis (W12), Fagus sylvatica–Rubus fruticosus (W14) and Fagus sylvatica–Deschampsia flexuosa (W15) woodlands. Here it forms part of a subcanopy of shade-tolerant trees being especially prominent in areas such as the New Forest and the chalk of south-east England. In Fagus–Mercurialis woodland, the T. baccata subcommunity (W12c) is found predominantly on south-facing steep calcareous slopes of the North Downs and some gaps in the Chilterns, with high exposure to wind and sun. Here T. baccata is a constant and it can be abundant as a lower tier of saplings beneath the Fagus or sometimes even break through into the canopy. It seems likely that in these woodlands, as in W13, T. baccata establishes in Juniperus-dominated scrub and the invasion of F. sylvatica is hampered by the shallow, impoverished soils of these steep, sun-baked slopes. Sometimes T. baccata occurs with Sorbus aria, although Fraxinus excelsior or Buxus sempervirens can be scarce companions but with little Crataegus monogyna or Corylus avellana. The field layer is usually a sparse mixture of Mercurialis perennis and Rubus fruticosus with a number of other species at low frequency. Thus there is a transition between W12c and T. baccata woodland proper (W13).
On the low base-status moist brown earths of the Fagus–Rubus woodland (W14), such as occasionally in the New Forest and sometimes in Surrey and Sussex, much more rarely in the Chilterns, and on the very base-poor, infertile soils of the Fagus-Deschampsia woodlands (W15) in the southern lowlands, the under-tier of Ilex aquifolium can be sporadically accompanied by scattered individuals of T. baccata, which may reach high enough to break almost through the canopy. These fall within Corine Code 9120/Palaearctic Habitat Classification 41.12 – Atlantic acidophilous beech forests (Quercinion robori-petraeae or Ilici Fagenion) (Interpretation Manual of European Union Habitats 1999). In a similar way, T. baccata can also be found in Quercus–Pteridium–Rubus woodlands (W10) on base-poor brown earths as a patchy lower tier beneath the oak, as in some New Forest stands.
As noted above, T. baccata woodlands are found in close association with grasslands and scrub which reflect grazing relaxation over the south-eastern chalk. Thus, yew stands can be found in mosaics and zonations with a complete sequence of vegetation types from close-cropped turf of Festuca ovina–Avenula pratensis (CG2), other swards dominated by coarse grasses to various kinds of Crataegus monogyna–Hedera helix scrub (W21).
In south-west Ireland, the Killarney woodlands contain extensive stands of yew dominating on bare Carboniferous Limestone pavement areas, and classified as facies of Taxus baccata of the Association Corylo × Fraxinetum (Order Fagetalia, Class Querco-Fagetea) (Ir. Pfl.) and also belong within Corine Code 8240 and Palaearctic Habitat Classification 62.3 (Interpretation Manual of European Union Habitats 1999). On more broken rocky terrain the shorter yew trees on the limestone outcrops are intermingled with taller Quercus (largely Q. petraea) in the intervening soil-filled hollows (Kelly 1981). Little grows under the yew canopy although Corylus avellana or Ilex aquifolium become frequent locally and Corylus can replace yew as the dominant. The field layer is sparse, the most constant species being Brachypodium sylvaticum, Fragaria vesca, Oxalis acetosella, Potentilla sterilis and Sanicula europaea. Unlike the yew woodlands of southern England, here there is a very dense moss cover, primarily Thamnium alopecurum with Eurhynchium striatum and Thuidium tamariscinum (Kelly 1981).
There is no direct equivalent of these species-poor mature stands in continental Europe. Delelis-Dusollier & Géhu (1972) and others put samples of fairly species-rich yew scrub from south-east England into a Roso-Sorbetum ariae which they took as the British equivalent to the Taxo-Prunetum mahalebis, a community found on chalk cliffs along the Seine. British yew woodland is best considered as part of the Fagion alliance in the Querco-Fagetea (see Kelly 1981; above), alongside the Fagus–Mercurialis woodland and the European communities such as the Taxo-Fagetum (e.g. Król 1978; Rodwell 1991; Saniga 2000).
From Portugal to the Crimea in the east, constrained by the need for a humid climate, T. baccata is usually a minor component of climax deciduous, coniferous or mixed woods (Tittensor 1980). For example, in northern Spain yew grows only as isolated individuals within the beech forest (Peñalba 1994); in central and eastern Europe yew is an understorey component of beech, or mixed forests of beech and conifers (e.g. Sokolowski 1921; Hoffmann 1958; Saniga 2000). Yew occurs in all ecological regions of Cephalanthero-Taxetum on shell-limestone or chalk (Król 1978). These communities are intermediate between Melico-Fagetum and Seslerio-Taxetum with the occurrence of Fraxinus excelsior and Tilia cordata. The community Seslerio-Taxetum is used to distinguish stands with yew from those with beech (Seslerio-Fagetum) – Moor 1952 cited in Król (1978). In the mountainous regions of Czechoslovakia, yew grows in fir and beech forests, and at lower elevations it is part of the oak and hornbeam forests (Futak et al. 1966). In Poland, yews grow amongst beeches on podsolic soils (Melico-Fagetum) and in a calcium-rich ash and alder carr (Circaeo-Alnetum) (Bojarska & Plitcha 1968, cited in Król 1978). Also in Poland, yew is found in Querco-Carpinetum medio-europaeum stands and with alder on fen soils (Carici elongatae-Alnetum).
Outside of Britain, T. baccata appears to form single-species stands only in the oceanic climates of the Crimea and Caucasus Mountains on a small scale (Pridnya 1984). Król (1978) records that, in the Caucasus, stands of pure yew or where yew dominates can be up to 10 ha with trees up to 32.5 m in height. These grade into forests formed by Fagus sylvatica, Carpinus spp., Fraxinus excelsior, Acer platanoides, A. campestre, A. pictum, Ulmus elliptica and Fagus orientalis mixed with Abies nordmaniana and Picea orientalis, with an understorey of Buxus sempervirens (Bugala 1978; Pridnya 1984). Purer yew woodlands are apparently being invaded by beech and ash. The islands of Corsica and Sardinia also have very local woodlands dominated by T. baccata, often with Ilex aquifolium and Buxus sempervirens (in Corsica), Sorbus aria and Mercurialis perennis (classified as Corine Code 9580 and Palaearctic Habitat Classification 42.A72 and 42.A73 in the Interpretation Manual of European Union Habitats 1999). These are thought to represent a senescent phase of the original beech or beech-fir wood. Similar T. baccata relicts, sometimes of small isolated formations, exist in the north and centre of Portugal.
IV. Response to biotic factors
Despite its poisonous properties, yew is very susceptible to browsing and bark stripping by rabbits, hares, deer and domestic animals such as sheep and sometimes even cattle (Elwes & Henry 1906; Watt 1926; Kelly 1975, 1981; Mitchell 1988, 1990a; Haeggström 1990; Sarmaja-Korjonen et al. 1991). Indeed, Kelly (1975) describes T. baccata as one of the most grazing-sensitive trees in the Irish Killarney woodlands. Despite this, yew is tolerant of repeated pruning (as demonstrated by its use in topiary) and is able to continue growth under severe browsing pressure (Tittensor 1980). Tabbush & White (1996) suggest that the swelling of many yew boles, with burr-like convoluted grain beneath, is probably the result of prolonged exposure to browsing. Welch & Haddow (1993) state that a feature of the whole genus is the almost total inability to develop fresh leader growth following damage; however, new leaders have been seen to form from buds just below the damaged tips 1–2 years after simulated browsing (Philip Perrin, pers. comm.). In the Burren of western Ireland, individual yew trees are restricted to prostrate dense shrubs, some fitting into solution hollows in the limestone pavement, primarily as a result of goat grazing; if the grazing is removed by exclosures, new vertical shoots produce rapid upwards growth. Trees can be killed by being scored by deer antlers, as has been observed in Killarney woodlands (Larner 1977), caused by Japanese sika deer (Cervus nippon).
Rabbit and deer grazing of seedlings and seed predation by rodents are known to be a major factor in lack of recruitment (Watt 1926; Williamson 1978; Kelly 1981; Mitchell 1988; Haeggström 1990; Hulme 1996; Hulme & Borelli 1999; Saniga 2000). The decline of yew in Denmark is attributed to widespread deforestation and selective felling followed by poor regeneration due to grazing (Svenning & Magård 1999). García et al. (2000) found that yew regenerated mostly under fleshy-fruited shrubs in the Sierra Nevada mountains of Spain (dominated by Juniperus communis, J. sabina, Berberis hispanica and Rosa spp.). The estimated density of the yew in the study plot was 288 individuals ha−1, more than 90% being juveniles (seedlings 82% and saplings 9%), which were mostly located under shrubs. This is dense regeneration, even by UK standards (see Hulme 1996). Saplings suffered serious herbivore damage when unprotected by shrubs. The high density of fleshy-fruited shrubs appears to be the key by (i) attracting frugivorous birds with a consequent accumulation of voided yew seeds beneath, (ii) maintaining a favourably moist and nutrient-rich microclimate (especially in the face of summer drought), and (iii) protecting against herbivores by the woody parts being relatively unpalatable with imbricate and often thorny branching. The shade-tolerant nature of the yew, of course, allows it to grow well in the heavy shade of the shrubs. In the stands studied, ungulates attacked 31.1% of marked juvenile yews. Damaged plants lost an average of 23 ± 3.6% (SE, n = 31) of their shoots to ungulates, damage reaching 67.7% in some individuals. Repeated herbivory was positively correlated with damage intensity which was negatively correlated with height and age. The risk of being grazed and the damage caused declined with increase in shrub cover and height.
Yew is shade tolerant even under dense shade (Lilpop 1931; Król 1978; Brzeziecki & Kienast 1994). However, yew establishment is hindered and saplings often die or show poor growth with few leaves when beneath the shade of beech (Czartoryski 1978; Król 1978; Pridnya 1984). In Slovakia, Saniga (2000) found that even though yew is very shade tolerant, a reduction in surrounding timber (dominated by Fagus sylvatica and Picea abies) by 18–20% volume improved the height growth of yew whereas a reduction in volume by 7–8% resulted in no growth above the control plots. He also found that with a good growth of Rubus sp., etc., on clear-cut sites, or with successful regeneration of beech, the conditions for the germination, survival and growth of yew were less favourable. Pridnya (1984) mentions that in the Caucasus broadleaf trees do not permit yew trees to grow because of their thick crowns which ‘offer mechanical interference’. In contrast, in areas with more continental climates, other trees (or rocky outcrops) protect against extremes of dryness or temperature, and yew is thus restricted to an understorey tree as in eastern Europe (Sokolowski 1921).
V. Response to environment
In optimum oceanic climates, T.baccata is gregarious, forming pure often even-aged stands with measured densities in Denmark of 1388 individuals ha−1, including 513 adults (≥ 30 mm diameter) (Svenning & Magård 1999). On the edges of yew stands at Kingley Vale, Sussex, where T. baccata is invading scrub, Williamson (1978) has described ‘families’ of yews where around four trees grow up through each juniper nurse plant, each some distance away from the next family; ‘The families may consist of one fat member, average girth 5 feet [1.5 m], two medium-sized, 3 feet [0.9 m], and one skinny member with as little as 2 inches girth [5 cm]’. Outside these stands, T. baccata is common as isolated trees on a wide range of soils. This is especially so in northern Britain and continental Europe where yew occurs mostly singly or as groups in the undergrowth. Extensive planting in the British Isles has led to the widespread occurrence of single or small groups of trees in gardens, towns and especially churchyards. Indeed in north-west Wales, yew is the commonest churchyard tree, constituting 33% of all trees (Denne 1987).
(b) performance in various habitats
Although, as noted above, saplings may do poorly or die under a canopy of beech, evidence suggests that once established yew can continue growing slowly in dense shade. Saplings can persist and grow very slowly for very many years before breaking through the beech canopy (Rodwell 1991). In Denmark, yew woodlands on glacial sand and clay have increased rapidly in radius at ≤ 3 m year−1 into beech forest (Svenning & Magård 1999). Nevertheless, growth of yew is slow compared to most other trees even under optimum conditions and after eliminating potential competitors. Consequently, even the oldest individuals do not attain considerable height. Height growth of 20–30 cm per year in the open is normal, trees reaching 4.5 m after 20 years. In Killarney woodlands, yew has a low canopy 6–14 m in height (Kelly 1981). There is a high energy investment in defensive mechanisms which increase resistance of wood against fungi and insect attacks.
Melzack (1979) found evidence of a geographical variation in yew growth in England with southern trees having the most vigorous shoots (in terms of leaf number) and the smallest leaves compared to those in the north-east. There was also a suggestion that trees in the south are bigger than those in the north-east. Watt (1926) pointed out that grazing can drastically affect net growth rates; under high grazing pressure from rabbits on the South Downs, he found one small yew just 18 cm high with approximately 55 growth rings.
Hulme (1996) shows that Castle Eden and Hawthorn Denes woodlands in Co. Durham have a truncated age-distribution − very few young trees − suggesting that over time there is declining recruitment but relatively constant mortality. This is consistent with the hypothesis that regeneration declines progressively as yew woodlands develop (Watt 1924, 1926). An important factor appears to be low levels of seedling recruitment (Williamson 1978; Kelly 1981; Hulme 1996). Hulme (1996) found 0–208 seedlings ha−1 in his Co. Durham woodlands. He suggests that this was partly because of high seed predation by small mammals (especially high beneath the shade of shrubs) but argues that lack of suitable microsites beneath scrub vegetation is an increasingly limiting factor in older yew woodlands as the scrub is progressively killed out.
Under a continental climate, growth is somewhat reduced as yew becomes more of an understorey tree. Many authors consider that in eastern Europe, yew grows better in shade, especially when young (e.g. Maly 1930). Traczyk (1953) suggested that yew will tolerate open conditions in such a continental climate providing it is on a fertile, moist soil.
Williamson's (1978) general figure for rates of trunk growth, as measured by width of annual rings, of 0.9–1.2 mm appears to be generally applicable in the British Isles. However, annual ring width has been found to vary from as little as 0.095 mm (a yew growing on thin peat and subsequently preserved in a bog of Co. Laois, Ireland; Adams 1905) to 3 mm or more on favourable soils (see Table 1). From empirical data, White (1998) suggests that maturity (judged by the cross-sectional area of successive rings reducing over time to reach a constant) is reached in 30 years for yews growing inside woodland, 40 years on poor ground or with some exposure, 55 years in churchyards and 60 years on an average garden or parkland site.
Table 1. Width of annual growth rings in the wood of Taxus baccata. These are made by direct measurement from cut trees or cores, or are averages inferred from widths or girths of trees of known age or trees measured at known dates
The rate of ring-width growth also depends on the age of the tree. As seen in Table 1, ring width slows dramatically once the yew reaches maturity, as it does in most trees. Newbould (1960) found ring widths of 0.13–2.54 mm in younger trees, falling to 0.051 mm in older trees. White (1998) gives a method for calculating the age of different species of tree from diameter based on empirical knowledge of how annual ring-width changes over time. White's data give an average ring-width for the first juvenile phase of growth (where ring width is approximately constant as diameter of the tree increases) of 3 mm year−1 for the first 30–60 years (least on most favourable sites), producing the rates of growth in Fig. 3. However, Tabbush & White (1996) found ring-widths on the dry, infertile soils of Kingley Vale, Sussex, to average 2.5 mm over the first 30 years, suggesting a slower rate of growth on these poor soils (see Fig. 3, curve b) compared to richer soils (curve a). Moreover, as trees enter senescence, with reduction in canopy size, growth rates will reduce still further. Meredith produced an empirical age-diameter relationship for yew using data from trees throughout the British Isles (data cited in Milner 1992; see Fig. 3, curve d).
Based primarily on measurements of ring widths and repeat measurements of old trees and trees of known planting date, Tabbush (1997) produced the following empirical formula for English churchyard yews (see Fig. 3, curve c):
Tree age = Girth (in cm)2/310
The data used extend to trees around 1000 years of age although, not unexpectedly, the scatter of points around the line (not shown in Fig. 3) increased with girth; overall, the curve accounts for 56.9% of the variation.
Moir (1999) used empirical measurements of yew growth from 12 trees growing at Hampton Court, London, on sand and gravel river terraces to produce a linear regression: age (in years) = 17.8 + 1.08 × girth (in cm); see Fig. 3. However, there is considerable variation between the age of a tree and its mean ring width or girth measurement. Again, Moir noted a great variation in growth rates such that trees in an avenue planted probably between 1890 and 1900 varied from 30 to 80 cm in diameter. The fairly high growth rates found here were attributed by Moir to the warmth created by the London conurbation.
Figure 3 shows the great variability between estimates of growth rates, dependent upon soil and other environmental factors. There is also a great deal of scatter around each line (not shown), suggesting very different growth rates even within the same site. Thus, a number of authors suggest that the girth has only limited use in estimating the age of a yew tree (e.g. Denne 1987; Moir 1999) and should be interpreted with reference to soil type, history of the canopy, exposure, etc.
Why is growth rate within one site variable? White (1998) points out that the yew is unusual in that it can return to juvenile rates of growth at almost any stage, perhaps stimulated by extra nutrition from branch layering and vigorous regeneration after catastrophic damage. Supporting this, a study of ring widths in a 335-year-old yew at Alice Holt Lodge, Surrey (T. Hindson pers. comm.) showed two distinct peaks of ring widths over 2.0 mm with the tree still producing rings over 1.0 mm wide after 300 years.
As is usually the case, growth rates (whether from direct measurements of ring widths or inferred from external measurements of the tree) also appear to vary according to where measurements are made on the trunk. The measurements by Lowe (1897) in Table 1 on trees 0.3 m in diameter and 75 years old were made at 3 feet (0.91 m) above ground level while the much larger average ring widths of Henslow (1889) were made ‘near the ground’ on trees 0.6 m diameter and 54 years old. Indeed, a number of records given by Lowe (1897) show initially much faster rates of growth near or at ground level but growth at ground level may appear then to stop or slow dramatically producing an ‘inverted’ tree shape. For example, the Ormiston Yew in Scotland in 1834 was 3.88 m in girth at ground level and 5.38 m at 1.5 m above.
The growth rates given above point to the antiquity of yews but the great variation makes it difficult to estimate accurately the age of any one individual. Certainly, ring counts of 400–500 are not uncommon (e.g. Bowman 1837; Bebber & Corona 1986; Parsapajouh et al. 1986). Ring counts cannot generally be used to date older trees reliably since these are invariably hollow; Mitchell (1972) stated that all yews over 4.5 m in girth are very likely to be hollow. Moreover, yew is renowned for missing or, at least, very narrow (and thus easily missed) rings. Lowe (1897) reported on a tree in Kew Museum that had 250 rings on one side and just 50 on the other. To help accurate counting, Moir (1999) points out that very narrow rings are usually more easily seen on the lobes/flutes common in yew trunks. Tabbush & White (1996) conclude that ‘the age of the oldest trees cannot be determined with certainty, but we can be confident that they are in excess of 2000 years old’.
There are other pitfalls in estimating age. As with other tree species, old senescent trees can produce new stems from remains of the former large trunk and appear as young trees. Grazing may also keep a tree artificially small. The yew, however, is also renowned for the apparent production of large trees from the fusing of several trunks. Lowe (1897) and Williamson (1978) give many instances of trunks with two or more centres surrounded by their own concentric growth. Williamson (1978) described a felled yew 37 cm in diameter which contained seven centres ranging from 21 to 53 cm in circumference, each 46 years old. Several possible causes for several centres have been considered. This may arise from a coalescence of several individuals (‘aggregates of individuals’Lowe 1897) either from dense new shoots from around the base of a dying trunk which become welded togther, or from a circle of seedlings, for example, around a juniper, which eventually coalesced. Chetan & Brueton (1994) argue that the latter is unlikely because such coalescence would normally result in male and female branches growing on the same tree which is not usually observed. Either way, yew is particularly prone to fuse with other stems and surround objects such as stones. Several centres may also result from the inclusion of branches into a trunk or, in very old trees, from the coalescence of roots originating at the bases of branches and growing down through the hollow centre of the tree (Tabbush & White 1996).
Lowe (1897) argued that the fusing of multiple stems/roots would make an individual appear much older than it really is. However, overall growth rate of the new trunk when young is unlikely to be much greater than a single tree of the same size. Indeed, Tabbush & White (1996) argue using rough calculations that the coalescing of seedlings would actually make little difference to age estimates. Providing the original circle of trees was not much larger than 1 m in diameter, the error in subsequent age determination of the resulting ‘tree’ would be minor.
(c) effect of frost, drought, etc
Taxus baccata is intolerant of severe and prolonged frost (Godw. Hist.; Brzeziecki & Kienast 1994; Skorupski & Luxton 1998) and icy winds (Bugala 1978). Elwes & Henry (1906) give examples of yew being damaged or killed by frosts during the severe winter of 1837–8 in the west of Scotland, and the death of many young shoots on trees at Burton-on-Trent in the winter of 1859–60. Till (1956) reported that needles from trees in southern Sweden showed a maximum winter resistance of −33 °C to −35 °C. Vegetative buds of the same trees showed a maximum resistance of −23 °C to −26 °C, whereas male reproductive buds were damaged below −21 °C to −23 °C. From the Austrian Alps, Pisek et al. (1967) reported that exposure to −21 °C for three hours resulted in 10–15% frost damage to needles and −23 °C resulted in 100% damage. Maximum tolerance to frost occurs in mid-winter (January), declining rapidly in early spring (Melzack & Watts 1982b; Brzeziecki & Kienast 1994) when tissue is vulnerable to severe frost conditions. In Britain, Melzack & Watts (1982b) found that, in midwinter, damage started at −13.4 °C whereas by March this had risen to −9.6 °C in the hardiest provenance (south England) and −1.9 °C in the most susceptible (north-east England). They suggested that this may explain the scarcity of native yew stands in the north-east. Sensitivity to frost in early spring is undoubtedly the crucial limiting factor in the northerly oceanic distribution in Europe of this species. Taxus × media is more resistant to low temperature in Poland than T. baccata, as also is its other parent, T. cuspidata (Bugala 1978).
The heat resistance of T. baccata (LT50, 30 min exposure) has two peaks, one in summer and one (the higher) in winter of 51 °C (Lange 1961).
Taxus baccata is tolerant of drought (Hättenschwiler & Körner 2000). Brzeziecki & Kienast (1994) rank T. baccata as 2 in a 1–5 scale where 1 is very tolerant of drought. The effects of drought are seen as needles older than 2 years turning yellow from their base upwards before falling (Strouts & Winter 1994). Experiments using double saw cuts to increase hydraulic resistance of the stem in ‘older male’ yews (Kartusch & Richter 1984) resulted in new needles above the cuts having more xeromorphic characteristics: reduced leaf area, thicker cuticle and a decreased number of stomata per unit area. A previous study (Richter 1974) had shown that these morphological changes were related to consistent differences (of up to1.5 MPa) in water potential between the treated and control shoots with the treated shoot frequently reaching a water potential of around −2.0 MPa. Parker (1971) observed an unusual lipid layer surrounding the tonoplast of the photosynthetic cells of several conifers, including T. baccata, which appeared to be related to drought resistance.
Watt (1924, 1926) surmised that wind is important in determining the distribution of yew on the South Downs as indicated by the extensive wind-shaping, the general limitation of the yew-woods to the valleys, and ‘from the greater and more rapid extension along sheltered slopes and to leeward of existing communities’ (Watt 1926). He also suggested that the real value of junipers is not just protection from grazing but the protection provided from exposed conditions. Certainly, at Kingley Vale, Sussex, salt winds from the sea (5 km away) turn exposed yew foliage brown-red, and trees on exposed slopes lean away from the south-west, the prevailing wind direction (Williamson 1978). Taxus baccata is sensitive to air-borne salt spray (Lemer 2001) and immersion in salt water (Buschbom 1968), winter shoots being 100% damaged by a short immersion in 23% NaCl or 23% CaCl2.
Owing to its thin bark, T. baccata is intolerant of fire (Hartzell 1995). Núñez-Regueira et al. (1997) found its branches (< 8 cm diameter complete with leaves, collected in the mountainous zone of north-west Spain) to be average in calorific value and flammability amongst 12 tree species tested (all the others being angiosperms). This suggests that the non-resinous T. baccata has low flammability compared to most gymnosperms.
VI. Structure and physiology
Taxus baccata is slow-growing, reaching 20(−28) m with a spreading, rounded or pyramidal canopy. The trunk is frequently very fluted and asymmetrical, and often with multiple stems. Branches are long, not whorled. Twigs are green and irregularly alternate. Even 4-year-old branches have parts still with an intact epidermis under which are live cork cells rich in chloroplasts (Hejnowicz 1978). Buds are very small, bud scales dark-brown, rounded, imbricate and closely appressed to the bud. When the bud opens, the bud scales unfold and remain at the base of the growing shoot, and on older branches persist as dry brownish scales. Lateral buds develop at irregular intervals, many of which remain dormant; epicormic buds are thus frequent along the upper sides of branches or on the stem which explains why T. baccata, unusually for gymnosperms, can withstand pruning back to the main trunk (Elwes & Henry 1906) and can produce leafy branches from old branches and trunks, and sometimes from stools. Yew occasionally produces burrs around the epicormic buds. The largest recorded is probably that seen by Howard (1944) from the Caucasus ‘measuring more than seven feet [2.1 m] in length and more than four feet [1.2 m] in depth, and weighing nearly a ton’. The bark is reddish-brown, thin, scaly with a rough and irregular outer surface due to exfoliation of the subereous layers as thin flat flakes. It is odourless and with a slightly bitter and astringent taste. The outer bark has a single periderm with 10–15 layers of phellem cells, the layers formed by alternating rows of thin- and thick-walled cells. The periderm does not have lenticels (Hejnowicz 1978). The phloem contains tangential bands of axial parenchyma with abundant calcium oxalate crystals embedded in the cell walls. Crystals of calcium oxalate are also in the phloem fibres (Hejnowicz 1978). Sieve cells are 150–200 µm in length. Di Sapio et al. (1997) give further details of phloem cell structure.
The wood of T. baccata lacks axial parenchyma (Hejnowicz 1978; Bell & Hemsley 2000) and resin canals although traumatic resin canals are formed when injured (Hejnowicz 1978). Woodpeckers (Picidae) are known to inflict damage on yew trees in Switzerland, resulting in traumatic xylem and phloem (Kučera 1972). The reddish heartwood is distinct from the pale sapwood. The early/late wood transition within each annual ring is gradual. The wood is hard, heavy and flexible, with a high density of 640 kg m−3 (Brzeziecki & Kienast 1994) to 700–800 kg m−3 (Schweingruber 1993). This is especially high for a gymnosperm, as dense as beech and even oak. The wood is described as one of the most durable of temperate trees (Eaton & Hale 1993). The tracheids have uniseriate pits and prominent spiral thickening; the latter may explain the great elasticity of yew-wood and its consequent use in bow making (Sporne 1974). The wood has a lower cellulose to lignin ratio (1.13), and an ash content (0.3%) which is 2–3 times lower than Pseudotsuga menziesii and Picea abies (Kučera 1998 cited in Lyubenova & Nedelchev 2001).
Narrow tracheid diameter makes for low permeability of gases –Eaton & Hale (1993) present data that show it is one of the least permeable softwoods (< 18.1 × 10−6 darcys). Not surprisingly, Redding (1971) classified yew as resistant to impregnation by creosote, requiring long treatment times and even then difficult to achieve better than 3–6 mm penetration. Retention of creosote was 336 kg m−3, comparable to ash but much more than oak (56 kg m−3) which was referred to as extremely resistant and virtually untreatable.
Like Fagus sylvatica, T. baccata has an extensive horizontal root system with a thick-felted mat of fine roots ramifying through the soil near the surface (Rodwell 1991). In some areas, notably on Carboniferous Limestone pavements in south-west Ireland, the roots of yew spread out over the rock surface and penetrate deeply into the fissures (Kelly 1981). On chalk, a similar situation can arise with the roots spreading over and penetrating into the surface of otherwise bare rock.
The needles are spirally attached but on lateral shoots are twisted more or less into two ranks; 1–3 (−4.5) cm long and 2–3 mm wide with (in England) a mean of 28 needles per shoot produced each year (Melzack 1979). They live for 4–8 years before falling (Bugala 1978; Szaniawski 1978). The lower epidermis and petiole have calcium oxalate crystals embedded in the cuticle. Four different long-chain esters have been isolated from the cuticular wax (Jetter et al. 2002). Stomata are found on the underside of leaves only, within two pale stomatiferous stripes but they are not arranged in lines within the stripes (Di Sapio et al. 1997). Stomata are sunken with elevated subsidiary cells. Stomatal density has been counted as 59 mm−2 (Di Sapio et al. 1997), 82–119 mm−2 (in 25 varieties of yew; 92 mm−2 in T. baccata proper; Dempsey & Hook 2000), and 115 mm−2 (Salisbury 1927). The petiole has no stomata.
Taxus has only endomycorrhizas (Levisohn 1954; Strullu 1978, 1991). Harley (1959) summarized the description of Pratt (1926), and Harley & Harley (1987) cite 13 references in their checklist. Successful beneficial inoculation with Glomus mosseae has been achieved by Gianinazzi et al. (1995). Sainz et al. (2000) have shown that there is a significant improvement in nursery stock growth of T. baccata after inoculation with either Acaulospora scrobiculata or Glomus deserticola. They also showed that neither of the latter endosymbionts protect the seedlings from infection by the root pathogen Phytophthora cinnamomi.
(c) perennation: reproduction
Taxus baccata is a long-lived phanerophyte. Old trees may show vegetative reproduction by branches touching the ground and producing adventitious roots (Suszka 1978), which can survive independently upon the demise of the main trunk. Taxus baccata will also sprout from the stump. Elwes & Henry (1906) record that yew never produces root suckers. Davis (1965) refers to large T. baccata trunks being built up of ‘coalescing suckers’ on steep slopes; however, Tabbush & White (1996) suggest that such new shoots match the description of Watt (1926) of the ‘pioneer’ seedlings or epicormic shoots fusing into one trunk. Fallen yew trees will continue to grow as long as some of the root plate is intact, with lateral branches becoming new leaders.
Taxus baccata can be propagated artificially through cuttings, seed, grafting or layering (Hartzell 1991). Experiments in Germany, where about 50% of all cuttings developed roots, have shown that clonal differences for rooting ability are high with 71.2% of the total variance explained by clone differences. No differences were found in the rooting ability between female and male clones by Schneck (1996) but Nandi et al. (1996) found that auxin treatment of cuttings induced higher rooting success in male trees (55%) compared to female (15%). The use of plastic tree shelters is not recommended for T. baccata as failures can occur owing to overheating.
2n = 24 (Dark 1932; Sax & Sax 1933; Lövkvist 1963, cited in Moore 1982). One chromosome has a terminal or subterminal attachment point, one is distinctly heterobrachial, while the others are more or less isobrachial (Sax & Sax 1933). Details of the meiotic stages of T. baccata are given by Dark (1932). Polyploids are not reported.
Mitchell (1998) compared sun grown and shade grown Taxus brevifolia with sun grown T. baccata. Above a PPFD (Photosynthetic Photon Flux Density) of 400 µmol m−2 s−1, both species had a similar maximum photosynthetic rate (4 µmol CO2 m−2 s−1 and 3 µmol CO2 m−2 s−1, respectively), both indicative of shade tolerant plants. Although both are shade tolerant, Mitchell (1998) found that T. baccata appeared to be better adapted to grow in full sunlight.
Taxus baccata can assimilate CO2 down to 2–3 klux (c. PPFD of 75 µmol m−2 s−1) (Pisek & Rehner 1958). The light compensation point varies from 175 lux in summer to 3.2 klux in February (Pavletic & Lieth 1958). Pisek et al. (1969) showed that the maximum summer rate of photosynthesis at 10 klux (c. PPFD of 300 µmol m−2 s−1) in T. baccata from the Austrian Alps is 7 mg CO2 gDW−1 h−1 whereas the maximum winter rate is 5.5 mg CO2 gDW−1 h−1. The optimum temperature range (90% of maximum or greater) they found was 14–25 °C. This optimum temperature for photosynthesis is higher than in other species of conifer. Summer temperature minimum for photosynthesis is −3 to −5 °C (Pisek et al. 1967) and winter minimum is −8 °C. The summer temperature maximum for photosynthesis is 38–41 °C (Pisek et al. 1968, 1969). A review (in English) of these data is given by Pisek & Larcher (1973).
Photosynthetic ability of T. baccata leaves declines with age, to 50% in 7-year-old needles compared to young needles (Szaniawski 1978).
(ii) Water relations
In his comparison of T. brevifolia and T. baccata (see above), Mitchell (1998) found no difference in stomatal conductance at a PPFD of 200 µmol m−2 s−1, but T. baccata showed significantly lower conductances at PPFD's of 400 and 900 µmol m−2 s−1 (stomatal conductances of 0.017 and 0.024 mol m−2 s−1 vs. 0.053 and 0.073 mol m−2 s−1 for T. brevifolia). He found no difference in stomatal density but there was a significantly higher number per leaf (4684 vs. 1604) as the leaf of T. baccata is larger.
Moir (1999) examined tree ring widths from yews at Hampton Court growing on sand and gravel and found that precipitation (particularly in the months February to July) had a positive effect on growth. Increased ring width was also positively related to late winter temperatures (January and February) and late autumn (October) temperatures, probably indicative of an extended growing season. High summer temperatures (particularly June) appeared to inhibit growth. Moir (1999) concluded that mild/wet winters combined with cool summers provide ideal growing conditions for yew, and that hot dry summers adversely affect yew growth. The yew chronology cross-matched well with those of Pinus, Fagus, Castanea, Carya, and to a lesser extent Quercus. In the Balkans, Lyubenova et al. (2001) found that 93% of ring width variation could be attributed to January–June precipitation of the current year and October–December precipitation of the previous year.
Ring width and maximal densities of rings in northern Iran have been correlated to February and March temperatures (Parsapajouh et al. 1986). Optimum temperature for meiosis in pollen mother cells is between 1 °C and 10 °C (Chira 1964). Temperatures above or below this, if lasting for several days, have a negative effect on meiosis causing partial or complete sterility (Vidaković 1991).
Fiedler et al. (1986) and Allen (1989) report that, compared to other conifers, T. baccata leaves have a high N, P, K, Mn, Zn, B and Mo content as well as low Al, Si and Fe content. With increasing needle age, the N, P, K, Mg and Mo content decreases (Fiedler et al. 1986). They also found that the content of Cu, Mn, Zn, Mo, Al and Si depends on the substrate; for example on acid silicate sites, old needles have very high Mn (up to 1.5%), Zn and Al levels. Nemec (1951) reported that, in Slovakia, the genus Taxus has high requirements with regard to mineral nutrition particularly for K, P and Ca.
Sun grown T. baccata leaves had a significantly higher nitrogen content (1.89% N and 0.22 mg N cm−2) compared to T. brevifolia leaves (1.21% N and 0.16 mg N cm−2) (Mitchell 1998).
Compared to angiosperm trees (Quercus robur, Fagus sylvatica, Castanea sativa, Acer pseudoplatanus and Sorbus aucuparia) growing in north-west Spain, T. baccata branches (< 8 cm diameter plus leaves) had high levels of Cd (up to 40 p.p.m.), Zn (41 p.p.m.), and Mn (up to 3074 p.p.m.) (Núñez-Regueira et al. 1997).
Several authors have investigated the composition of T. baccata arils. Herrera (1987) found arils (with seed) from Spain were typically 764 mg fresh weight (10.9 mm length × 10.3 mm width – larger than the average, see introduction), with a dry pulp of 158.7 mg. Percentage moisture of the aril and seed together was 66.3% while Suszka (1978) found that fresh seed in Poland had 31.1% moisture, and Hegnauer (1962, cited in Bartkowiak 1978) measured the aril alone to have 80% moisture. The dry pulp consists of 0.2% lipids, 2.3% protein, 2.6% fibre, 93.5% non-structural carbohydrate and 1.4% ash (Herrera 1987). The following mineral contents were also found (in g kg−1): Ca, 0.2; Mg, 0.1; P, 0.4; K, 6.0; Na, 0.2, and (in mg kg−1) Fe, 25; Mn, 1; Zn, 5; Cu, 1. Energy content has been variously measured: dry aril pulp, 3.88 kcal g−1 (Hulme 1997); ripe arils 5.07 kcal g−1; embryo and cotyledons 8.41 kcal g−1 (Smal & Fairley 1980b).
(v) Other aspects
Taxus baccata is tolerant of air pollution (Mez 1963) with a high resistance to SO2 (Białobok 1978), though concentrations higher than 50 mg m−3 damaged needles. Szaniawski (1978) reviewed the literature and commented that T. baccata is more tolerant of smoke from copper foundries and to hydrogen fluoride than Pinus or Abies species. Emissions of heavy metals in Germany have been effectively monitored using T. baccata needles as a bioindicator (Höllwarth 1984). Barker & Ashenden (1993) showed that acid fog treatment of pH 2.5 significantly increased T. baccata needle damage relative to pH 3.5–5.5.
Hättenschwiler & Körner (2000) and Hättenschwiler (2001) exposed seedlings of six major European temperate forest tree species, including T. baccata, to elevated levels of 500 p.p.m. CO2 and 660 p.p.m. CO2 in the understorey of an old-growth forest in Switzerland. Seedlings were exposed to a known quantum flux density (QFD) ranging from 0.36 to 2.16 mol m−2 day−1 (0.8% to 4.8% of full sun). Biomass production in the deep-shade tolerant species (Fagus sylvatica and T. baccata) increased by 73% and 37%, respectively, under elevated CO2 in low light microsites but was not signi-ficantly different among CO2 treatments in high light microsites. Thus, these two species are likely to be at an advantage at higher carbon dioxide concentrations but may lose their advantage in only slightly higher light levels which still represent shaded conditions. As CO2 concentration rises, germination, recruitment and survival of these two species in deep shade conditions are likely to improve.
(f) biochemical data
All parts of the tree except the aril contain the alkaloid taxin (Vidaković 1991).
There is a very large literature on several biochemistry aspects, particularly on the anticancer properties of the taxanes. A short account is given here.
An overview of the history, chemistry, supply, biology and clinical aspects is given in Suffness (1995). Originally extracted from the bark of the Pacific yew (T. brevifolia), taxol (paclitaxel) can now be synthesized from a taxol congener (10-deacetylbaccatin III) which can be extracted at approximately 10 times the quantity per unit weight from the leaves of T. baccata and other Taxus species than from T. brevifolia bark (Denis 1988). Thus the final taxol is ‘semisynthesized’ from the taxane core (Nicolaou et al. 1996). Taxol has also been synthesized de novo (Nicolaou et al. 1996). Taxotere (docetaxel) is a semisynthetic taxane also derived from 10-deacetylbaccatin III from the needles of T. baccata (Vaishampayan et al. 1999). A very large number of taxanes and their derivatives have been isolated from yew trees including T. baccata (Parmar et al. 1999). Guéritte (2001) has reviewed recent developments in taxoid chemistry and also has given the structure of these taxoids, as have the other papers cited in this paragraph.
Vesela et al. (1999) found the highest concentration for all taxanes examined in T. baccata in Czechoslovakia to be in October and the lowest in January. Hook et al. (1999) examined French and Irish trees of T. baccata and concluded that there was not a consistent maximum annual peak of taxoids as it appeared to vary with location, variety and taxoid type. The taxane content of the needles, stems and clippings of 92 cultivars (mostly T. baccata) obtained from nurseries in the USA is reported by Elsohly et al. (1997). The taxane content of the leaves, stems and clippings in the studied cultivars ranged from 0.0135–0.1471% (w/w), 0.0055–0.0462% and 0.0121–0.1183%, respectively. 10-deacetylbaccatin III and taxol were the most abundant.
Tissue culture of Taxus species is now an alternative source for taxoid production (e.g. Jaziri et al. 1996). However, cultures based on T. baccata varieties generally showed weak biological activity in the tubulin test compared to T. media varieties (Parc et al. 2002). Tissue culture of T. baccata is being used in an attempt to increase taxane production either by purification of important synthetic pathway enzymes, e.g. geranylgeranyl diphosphate synthase (Laskaris et al. 2000), or by isolation of biosynthetic genes (Walker & Croteau 2001).
The seeds and foliage of T. baccata are rich in toxic alkaloids (Miller 1980), all parts being poisonous apart from the aril. They contain taxin(e), a complex mixture of alkaloids that is rapidly absorbed from the digestive tract and interferes with the action of the heart (Cooper & Johnson 1984). All parts apart from the arils contain a cyanogenic glycoside–taxiphyllin (Cooper & Johnson 1984; Barnea et al. 1993), but the poisonous properties of yew are primarily due to the alkaloid taxine (Clarke et al. 1981). A review of the toxic properties of these alkaloids in T. baccata has been given by Wilson et al. (2001), where they concluded that the chief action of the taxines in humans and other mammals is on cardiac myocytes with resultant heart failure and death. Shanker et al. (2002) have also implicated toxic effects involving the liver.
Erdemoglu & Sener (2001) found that ethanol extracts of heartwood of yew collected in Turkey exhibited a significant activity against some gram-negative bacteria but not against gram-positive bacteria. It also showed antifungal properties against five species of fungi of medical importance.
‘Flower’ buds form in the second half of the summer. They open the following spring (February/March in western Europe and into April in the east and north) and are thus one of the earliest British trees to flower with Ulmus and Corylus. In western Europe, Richard (1985) found that yew strobili open 100 days after mean daily temperatures fall below 10 °C in autumn. Fertilization occurs 6–8 weeks after pollination (Hejnowicz 1978; Pennell & Bell 1986). ‘Fruits’ ripen and seeds are shed primarily in August to October in mainland Europe and north-eastern USA (USDA 1974; Bugala 1978) extending into November in Russia (Sukachev 1928; Wyman 1947) and southern Ireland (Smal & Fairley 1980b).
VIII. Reproductive and seed characters
(a) reproductive biology
Pollination anemophilous although honeybees have been observed collecting pollen. When fully mature, the nucellus secretes a sugary fluid which accumulates as a drop at the orifice of the micropyle (Hejnowicz 1978). This pollination drop serves to trap wind-borne pollen. Subsequent reabsorption of the drop draws the pollen down to the surface of the nucellus where it germinates (Ruguzov et al. 1992). The male gametes are non-motile. Seeds mature within 1 year.
Svenning & Magård (1999) observed that production of reproductive organs increased strongly with tree diameter in Denmark and also with canopy openness. In adult trees, 62% of female trees and 29% of males produced reproductive organs in the open, while in shade, 19% of females and 10% of males produced them. Watt (1926) noted that yew does not produce ‘fruit’ under a closed beech canopy.
Taxus baccata is dioecious. Sex ratios have been found to vary from 44% female in Kingley Vale, Sussex (Williamson 1978), and 47.8% female on an island in Loch Lomond, Scotland (Pilkington et al. 1994), to 67% female in the Caucasus Mountains (Pridnya 1984) and 70% female in the Sierra Nevada Mountains of Spain (García et al. 2000). Change in sex of an individual tree is possible. The Irish Yew (T. baccata Fastigiata) is derived from cuttings from a female tree transplanted into Florence Court, Co. Fermanagh, Ireland; in 1927, however, male Irish yews were discovered in Sussex (Hartzell 1991). Bowman (1837) remarked on a 60-year-old tree in Shropshire (almost 7 m in girth) which had one female branch on an otherwise male tree. As in this case, such examples are normally of a single branch carrying reproductive structures of a different sex from the rest of the tree. However, Elwes & Henry (1906) reported on a tree near Stuttgart that bore male and female structures irregularly over the whole tree, each on separate twigs. Very occasionally both sexes may be found on the same branch. In the Caucasus Mountains, Pridnya (1984) found 1% of trees to be bisexual.
The Japanese yew, T. cuspidata Siebold & Zuccarini, hybridizes freely with T. baccata to produce a variable hybrid Taxus × media Rehder. The cross was first made by T.D. Hatfield at Hunnewell Pinetum, Massachusetts, around 1900. A chemotaxonomical study of the genus, based on the concentration of six neutral taxanes in its needles, showed some separation of the species T. baccata, T. cuspidata and T. × media, but did not indicate an intermediate position for T. × media, but rather a heterosis effect for the hybrid (van Rozendaal et al. 1999).
(c) seed production and dispersal
Seed production starts at 30–35 years old (USDA 1948; Vidaković 1991) in single trees and open stands, and at 70–120 years old in dense eastern European stands (Brzeziecki & Kienast 1994; Lyubenova & Nedelchev 2001). In Britain, prolific ‘fruit’ is produced in most years (Fuller 1982; Snow & Snow 1988) although in less optimal eastern European areas good seed crops occur every 2–3 years (Brzeziecki & Kienast 1994). In the Killarney yew woodland, the number of part and complete ‘fruits’ falling was estimated at 2.6 and 6.2 million ha−1 in the two years, respectively, beginning autumn 1975 (Smal & Fairley 1980b). This corresponds to 96.4 and 308.3 kg ha−1 with a total energy content of 0.56 and 1.87 m kcal ha−1, respectively. The authors estimated that birds removed 35.1% and 43.3% of the weight of arils, in each of the two years. Smal & Fairley (1980b) should be consulted for a breakdown of these figures into different ‘fruit’ categories.
The average fresh weight of T. baccata seeds has been reported as 57 mg (n = 20), Lichfield, Staffordshire; 64 mg (n = 50), Keele University, Staffordshire; 64 mg (n = 140), Box Hill, Surrey; 69.9 mg, Durham (Hulme & Borelli 1999). Melzack & Watts (1982a) found the average seed weight from six stations throughout England to be 56.5 mg. They also found that seeds from the south and west of Britain, with a more oceanic climate, were significantly heavier than seeds from drier more central and eastern sources. Thus weights ranged from 69.5 7.3 mg (SD, n = 30) for Butser Hill, south Sussex to 44.9 6.6 mg (n = 30) from Overton Hall, Derbyshire. Also there was some evidence (but not significant) that seed weight increased from the south at Butser Hill to the north at Yew Barrow, Cumbria (60.7 7.1 mg, n = 30).
A similar trend in average seed weights appears across continental Europe, from the oceanic west (77 mg in the Netherlands (Detz & Kemperman 1968) and 69.9 mg on the Iberian peninsula (Herrera 1987)) to the continental climate of Poland (43–59 mg; Suszka 1978; Brzeziecki & Kienast 1994). This raises the question of why seed in Britain is, on average, lighter than in western continental Europe. Seed from T. baccata grown in the USA range from 59 to 76 mg (Heit 1969; USDA 1974).
The main agent of seed dispersal is birds which consume the aril and seed, and disgorge the seeds or pass them in their faeces (Bartkowiak 1978; Fuller 1982; Snow & Snow 1988; Hulme 1996). Most birds excrete the seed although Ridley (1930, cited in Bartkowiak 1978) found that the mistle thrush (Turdus viscivorus) eats large quantities of the ‘fruit’ and sometime later they ‘spit out a red mass of partly digested pulp with brown yew seeds’. Up to 23 seeds were found in one spit.
Bartkowiak (1978) proposed that the birds that eat yew ‘fruit’ are limited to those that can swallow such a big seed, and indeed the ‘fruit’ is eaten primarily by the thrushes (especially the blackbird Turdus merula, mistle thrush T. viscivorus, song thrush T. philomelos, fieldfare T. pilaris, redwing T. iliacus and ring ouzel T. torquatus), starling (Sturnus vulgaris), waxwing (Bombycilla garrulus) and jay (Garrulus glandarius) (Creutz 1952; Bartkowiak 1970, 1978; Snow & Snow 1988; García et al. 2000). However, Bartkowiak (1970) also reports that in the Kórnik Arboretum, Poland, seedlings are found at the root collar of trees (and from cracks in trunks and even walls) resulting from seeds stored up to 50–70 m away from the parent tree by the nuthatch (Sitta europaea).
Vogler (1904) attributed the abundance of yew on cliffs in Switzerland (especially on rock ledges used as bird roosts) partly to the seeds spread by birds that can travel long distances. This leads to the occurrence of isolated yews on cliffs remote from other yew populations. This interpretation may not always be true because isolated trees can be remnants of previously extensive woodland.
In Oxford, Barnea et al. (1993) found that blackbirds and starlings spent a comparatively short-time visiting yew trees (an average of 1.3 ± 0.6 min (SE, n = 12)) and ate 4.3 ± 2.8 (n = 15) ‘fruits’ on each visit, compared to visits to hawthorn of 3.0 ± 2.4 min (n = 26) and a meal size of 5.7 ± 4.4 (n = 34) ‘fruits’. Unlike the other fruits tested (holly, ivy and hawthorn) which contained cyanogenic glycosides only in the fleshy part of the fruit, yew was found to have these only in the seed coat and not the aril (no species tested had them in the seed itself). Poisons in the fleshy parts are often explained (e.g. Sorensen 1984) as an adaptation to prevent birds consuming too many seeds at once, which, with the short visit time, will ensure better seed dispersal as only few seeds will be deposited at one site at one time. However, visit time and meal size are smaller for yew than hawthorn, despite the lack of poisons in the aril, so other factors must limit feeding apart from toxins in the food. It seems rather that the toxins in the seedcoat are to prevent eating of the seed itself by wood pigeons and crows (see section IX). Small meal size may also explain why so many yew seeds are left on the tree to fall beneath the canopy.
Bartkowiak (1978) in reviewing the European literature found a number of records of squirrels dispersing yew seeds by eating the aril and seed, and regurgitating the seed. In the Andalusian highlands, Hulme (1997) found that two species of ant, Cataglyphis velox (Sants.) and Aphaenogaster iberica (Em.), were responsible for the removal of a few percent of seeds from traps on the ground: they consumed the aril but not the seed and so contributed to small-scale redistribution of seeds. Seeds taken from the ground by rodents are hoarded and may germinate. However, Hulme & Borelli (1999) found that shrubs act as sinks for the accumulation of bird-dispersed seeds (preferential deposition), provide a suitable microclimate, and protection from large herbivores but also harbour high rodent pressure on seeds (indeed a large number of empty seed cases are common under yew in the UK).
(d) viability of seeds: germination
Seed viability can be high, approaching 100% (USDA 1948) but germination rates of 50–70% are more normal (USDA 1974) although Heit (1969) achieved 95% germination over a 3-year period. The viability of T. baccata seeds can be maintained artificially for up to 4 years by storing in damp sand or peat at ‘low temperatures’ (USDA 1974). Yew seed can also be stored for 5–6 years by drying at room temp to c. 10% moisture content and then storing at 1–2 °C (Heit 1968, 1969) or even lower at −3 °C (Suszka 1985).
The strong but variable dormancy is attributed to the small immature embryo present when the seed is mature. Certainly, the seed coat is not impermeable to water, artificial scarification with sulphuric acid does not enhance germination, and neither alternating temperatures nor short periods (20–30 days) of moist pre-chilling increase germination (although Williamson (1978) suggests that frost is necessary to break down the seed case and overcome dormancy). Various authors (e.g. Heit 1968 and ISTA 1966) have stated that 2–9 months of pre-chilling at low temperatures (typically 3–5 °C) is needed prior to germination. Suszka (1985), however, maintained that a cold-only stratification by itself does not appreciably promote germination even after 4 years of treatment, though the seeds remain viable. Indeed, Suszka (1985), and a number of researchers before him (e.g. Mitiska 1954; Heit 1968, 1969), found that a warm moist period (e.g. 15–20 °C for 90–210 days) before chilling (e.g. 60–120 days at 2–5 °C) is essential. Suszka (1985) found using Polish seed that under such treatment the small embryos almost doubled in length and changed from white to green during the initial warm phase. Suszka (1985) also recommends that a fluctuating temperature in the pre-chilling warm period is essential for high subsequent germination. Such a treatment appears to simulate seed in the soil passing through a summer and winter, thus explaining why most seeds germinate in the second year.
Page-Degivry & Garello (1973), working with T. baccata Fastigiata, found that embryo dormancy was eliminated when cultured in vitro in a liquid nutrient medium. They found no morphological development; rather germination was produced by a leaching of water-soluble germination inhibitors. However, distilled water alone did not remove dormancy; this was achieved only when sucrose was added to the medium and was further enhanced by the addition of Ca or K ions.
Natural germination of T. baccata is improved if the seeds have passed through the digestive tract of birds (USDA 1948; Suszka 1978). Indeed, Williamson (1978) planted 600 seeds in mouse-proof cages at Kingley Vale, Sussex, and obtained just 2% germination. However, treatments with hot water or sulphuric acid to break down the seed coat do not result in higher germination (Suszka 1978). Heit (1969) surmised that it was not the bird's digestive juices that helped germination, but the removal of the aril which is beneficial: he found that in pot trials there was less decay of seeds if the aril was removed, resulting in 4–6% germination with arils compared to 27–28% without.
Once dormancy is broken, a temperature range of 13–16 °C has been found optimal for maximum germination, which gives higher germination than 20 °C (Suszka 1978). However, alternating temperatures speed germination: ideal night/day temperatures have been variously recommended as 20/30 °C (ISTA 1966), 10/20 °C in Belgium (Devillez 1978) and 3/20 °C for Polish seeds (Suszka 1978). Germination is not stimulated by light. Vogler (1904) suggests that yew can germinate and grow on fully exposed bare rock as long as water is not limiting, which explains why yew is frequently found in cliff environments in Europe.
Melzack & Watts (1982a) found that seeds from the oceanic west of England were larger than those from the drier east. However, in laboratory trials there was no significant relationship between seed weight and germination rate or seed weight and seedling weight two months after germination. The only suggestion of the effect of seed weight is that the lightest seed from Overton Hall, Derbyshire (averaging 44.9 mg), produced the most erratic germination pattern over time with two peaks of germination.
(e) seedling morphology
Germination is epigeal. Seeds have two cotyledons with rounded apices, up to 2 cm long with stomata bands above (Vidaković 1991). Three to four opposite pairs of leaves are produced above the cotyledons, followed on vigorous plants by a few alternate leaves around the terminal bud. By end of first season, the seedling is usually 2–8 cm high. A strong tap root with a few laterals develops in the first year (Elwes & Henry 1906). Subsequent growth for the next 4–5 years is slow, often less than 2.5 cm annually. Seedling development is shown in Fig. 4.
IX. Herbivory and disease
(a) animal feeders or parasites
Despite popular opinion and the views of early writers (e.g. Tansley & Rankin 1911), yew is not invariably toxic to larger mammals (Clarke et al. 1981). Herbivory of leaves, twigs and bark is recorded by a range of mammals. The trees are generally susceptible to browsing by roe deer (Capreolus capreolus) (Tittensor 1980) and in south-west Finland can be thus ‘jeopardized’ (Klötzli 1965). In Norway, roe deer grazing on yew in winter was highest when bilberry (Vaccinium myrtillus) was in low density or snow covered. Observations suggest that roe deer may be the main factor affecting yew survival in areas where other preferred feeding plants are not available (Mysterud & Ostbye 1995). Dickson (1994) records fallow deer (Dama vulgaris) feeding on yew in Scotland but Dietz et al. (1994) record yew poisoning of fallow deer in Denmark. Deer have also been noted as causing a persistent problem by browsing on Taxus species in nurseries in north-east USA (Lemieux et al. 2000).
Hares and rabbits appear to browse T. baccata regularly without harm (Lowe 1897; Watt 1926; Williamson 1978). Grey squirrels (Sciurus carolinensis) will also browse on bark and twigs (Tittensor 1980) and eat arils (although the seed are later regurgitated; Bartkowiak 1978). Williamson (1978) comments on seeing foxes (Vulpes vulpes) and badgers (Meles meles) eating arils from the ground in Kingley Vale, Sussex, the former probably regurgitating the seed, the badger crunching the whole ‘fruit’ up. Bartkowiak (1978) records large quantities of seeds and undigested arils in badger faeces. He also mentions martens (Martes martes) eating T. baccata‘fruit’ in the Caucasus.
Wild Barbary macaques (Macaca sylvanus) in Morocco forage for T. baccata‘fruit’ (Mehlman 1988).
Small mammals also play a part in the herbivory of yew, especially of the seeds (Tittensor 1980; Hulme & Borelli 1999). Rates of yew seed predation may be very high particularly during the fruiting period of yew when it is the main diet of the rodents (> 60% of seed fall predated in Co. Durham; Hulme 1996). Yew woods support higher densities of rodents (principally wood mouse Apodemus sylvaticus and bank vole Clethrionomys glareolus) than comparable deciduous woodlands (Smal & Fairley 1980a). Watt (1924, 1926) and Williamson (1978) point out that although plenty of seeds were found in gaps between and under yews on the Sussex Downs, the vast majority had been opened and eaten by small mammals (or birds, presumed to be tits). In Durham woodlands, Hulme & Borelli (1999) found that rodents foraged more beneath the cover of shrubs (c. 60% removal of seeds and ‘fruit’) than in open areas (38% removal). They identified Apodemus sylvaticus and Clethrionomys glareolus as being principal seed consumers using selective exclosures. Unidentified vertebrate seed predators (probably birds, rabbits and/or squirrels) also appeared to be significant seed removers in three of the five woodlands examined. Rates of removal differed among the three tree species studied, increasing in the order Fraxinus excelsior < T. baccata < Ulmus glabra, but were not related to seed mass or seed density. The larger seeds of yew benefited little from seed burial whereas, by comparison, encounter of the smaller Ulmus seeds fell by almost two-thirds. Hulme & Borelli conclude that the opportunity to escape seed predation via burial or reduced seed density is limited, and that while seed predators may influence regeneration of small-seeded Ulmus, it probably plays a lesser role in the dynamics of T. baccata and Fraxinus.
A similar study done by Hulme (1997) in the Andalusian highlands, south-east Spain showed losses of 87% of seeds over three day periods. Rodents, primarily Apodemus sylvaticus, were the main post-dispersal removers of 73% of seeds. A greater proportion of bare seed (56.8%) was removed compared to intact ‘fruit’ (aril and seed) (49.5%), and indeed on a ‘number of occasions’ rodents removed the toxic seeds and left the empty energy-rich arils behind.
Cattle and especially horses can suffer quick and fatal effects from eating yew twigs or bark (Lowe 1897; Watt 1926; Knowles 1949; Williamson 1978; Kite et al. 2000) and will therefore produce comparatively little browsing damage. Cattle have been observed by Haeggström (1990) not to eat T. baccata in wooded meadows in south-west Finland. However, yew has been used as an animal fodder in continental Europe (Williamson 1978) and in Asia (USDA 1948), and cattle, sheep (Lowe 1897) and horses (Lowe 1897; Elwes & Henry 1906; Stählin 1944 cited in Haeggström 1990) appear to build an immunity to yew if regularly fed small quantities. In the mountains of Hanover and Hesse in Germany the peasants fed their cattle with branches of yew during the winter, beginning with small quantities and gradually increasing the amount (Williamson 1978). Small quantities and habitual access seems to be the key because there are many records of cattle freely browsing in pastures with yews (but with plenty of other fodder available) which suffer no ill-effects (Lowe 1897). Animals let into fields with yew in winter have been recorded as dying presumably because yew was the only green food readily available. Access to cut branches has led to many domestic animal deaths. Foliage is reputed to be even more poisonous when wilted or dried (Elwes & Henry 1906; Williamson 1978) but according to Cooper & Johnson (1984) they are ‘as toxic as the fresh plant’.
The lethal dose of T. baccata leaves is given by Völker 1950 (cited in Clarke et al. 1981) as 100–250 g for sheep, 500 g for cattle, 100–200 g for horses and pigs, and 30 g for a dog. Cornevin (1892, cited in Clarke et al. 1981) gives oral lethal doses (in grams of leaves per kg of animal) as: donkey and mule, 1.6; horse, 2; pigs, 3; domestic dog, 8; sheep and cattle, 10; goat, 12; rabbit, 20. The greater resistance of the ruminant is probably due to the dilution of the ingested yew with the contents of the rumen (Clarke et al. 1981). Panter et al. (1993) found an estimated dose of 0.36–0.70 g kg−1 of T. baccata clippings killed 35 cattle in western USA; this is much less than the 10 g kg−1 quoted above, but was a one-off exposure to wilted foliage.
Yew is poisonous to humans, causing heart failure (Clarke et al. 1981; Von der Werth & Murphy 1994; Willaert et al. 2002). Wine from hip-flasks made of yew wood was said by Pliny to poison travellers (Woods & Calnan 1976) and ‘miliary eruptions’ were reported from sleeping under the tree (Piffard 1881). Various authors report skin irritation and dermatitis caused by yew wood (Senear 1933; Long 1938; Wahl et al. 1992). Krenzelok et al. (1998) investigated 11 197 records of yew poisoning (from all Taxus species) in humans (96.4% in children less than 12 years old) and found no deaths. Van Ingen et al. (1992) reported that to their knowledge only 10 authenticated cases of fatal human poisoning by T. baccata have been recorded in the scientific literature in the previous 31 years, and these were all deliberate. Fifty to one hundred grams of chopped leaves is considered fatal to adults. However, adults and children of the Upper Valley of the Serchio River of Central Italy, who have a traditional diet of wild collected food, regularly eat T. baccata arils as a snack but often with the seeds as well (which Lowe (1897) says are almond flavoured). Pieroni (1999) suggests that the seeds are voided before they can be broken down to liberate the toxic alkaloids.
Many birds are known from yew woodland (see Williamson & Williamson 1973; Batten 1976; Williamson 1978). Snow & Snow (1988) record 18 species of birds, one third of them in the thrush family, that eat yew arils. Song thrushes (Turdus philomelos) and mistle thrushes (T. viscivorus) are the main seed takers in southern England (Hartley 1954; Williamson 1978). In autumn and winter large groups of thrushes feed on yew arils. In late summer, flocks of mistle thrushes descend on some yews and defend the aril crop which may not be eaten until late in winter (Fuller 1982). In October large flocks of redwings (Turdus musicus) and fieldfares (T. pilaris) join the thrushes until mid winter when aril stocks are depleted (Tittensor 1980). Williamson (1978) records seeing fieldfares remove up to 30 ‘fruits’ at one time. After mid winter, blackbirds (Turdus merula) are the commonest thrushes. Arils are also taken by ring ouzel (T. torquatus), robin (Erithacus rubecula), blackcap (Sylvia atricapilla), starling (Sturnus vulgaris), nuthatch (Sitta europaea) and pheasant (Phasianus colchicus) (Heymer 1966; Bartkowiak 1978; Bugala 1978; Tittensor 1980; Barnea et al. 1993; García et al. 2000). Starlings eat ‘fruit’ in Britain and Saxony (Heymer 1966) but do not seem to in eastern Europe (Hungary, Slovakia) (Szijj 1957; Havlin & Folk 1965). In the Andalusian highlands, Hulme (1997) found that most seeds were taken from the ground (and eaten) by small rodents but approximately an extra 14% of seeds were removed when animals larger than rodents had access; this was attributed in part to birds, particularly chaffinches (Fringilla coelebs).
The greenfinch (Carduelis chloris) is the main seed predator, but seeds are also eaten by great tit (Parus major), bullfinch (Pyrrhula pyrrhula), hawfinch (Coccothraustes coccothraustes), nuthatch, green woodpecker (Picus viridis) and great spotted woodpecker (Dendrocopos major), and occasionally by blue tit (Parus caeruleus) and marsh tit (P. palustris) (Tittensor 1980; Snow & Snow 1988). Greenfinches, however, undoubtedly destroy a small fraction of the yew crop compared to that taken by the main seed dispersers (Snow & Snow 1988). Greenfinches mandibulate the ‘fruit’ until the aril falls, then strip off the brown seed coat (which contains cyanogenic glycosides; Barnea et al. 1993) before eating the rest of the seed. Barnea et al. (1993) therefore suggest that the poison is concentrated in the seed coat which explains why woodpigeon (Columba palumbus) and some of the larger crows (which grind the whole seed in the crop) do not eat the ‘fruit’. Tits on the whole hold the seeds in their feet and hammer them open with their beak. Nuthatch and woodpeckers take the seed to a neighbouring tree, wedge it in a crack and open it. Nuthatches rub the aril off on the tree bark before wedging the seed into the bark. Chaffinches feed on the fragments of endosperms dropped by greenfinches (Snow & Snow 1988).
Lowe (1897) states that birds of all kinds are poisoned by eating yew leaves. Pheasants are noted as being prone to this (Tegetmeier 1896; Jordan 1964) and Williamson (1978) records that pheasants around Kingley Vale, Sussex, that live their lives within the yew trees, tend to be smaller than average and ‘are not particularly tasty to eat’. Poisoning by yew leaves is known in domestic poultry (Orr 1952).
Young individuals of two Mediterranean tortoises (Testudo hermanni boettgeri and T. marginata) have died after ingesting dried yew branches (Wiechert et al. 2000).
The yew big bud mite (Cecidophyopsis psilaspis Nalepa: Eriophyidae) is considered a serious pest of yew in northern and central Europe. Feeding the mites results in abnormal elongation and swelling of bud scales, leading to extensive and chronic bud mortality throughout the canopy and an irregular branching pattern, producing an asymmetrical crown (Duncan et al. 1997). Skorupski & Luxton (1998) list 199 species of mesostigmatid mites from 19 Orders sampled from T. baccata in England and Wales. These are mostly free-living predators although many ectoparasitic forms are known and some may consume carrion or microbial tissue. It is unlikely that any of the species are strictly confined to T. baccata. There is a distinct difference in species associated with ‘solitary yews’ (39 species found solely within the environment of solitary yews) and those in stands (‘forest yews’, 25 species of mites). The most frequently represented Orders were Parasitidae, Laelapidae and Zerconidae. The most widespread species were Pergamasus crassipes, Proctolaelaps pygmaeus, Veigaia nemorensis and V. planicola. The most frequent species (appearing in > 20% of samples) were Veigaia nemorensis, V. planicola, Uropoda minima, Paragamasus lapponicus, Proctolaelaps pygmaeus, and Cosmolaelaps claviger.
False spider mite Pentamerismus taxi (Haller) (Tenuipalpidae) has been found at the base of yew needles, being originally described from yews in Switzerland (and also recorded on the southern coast of the Crimea, see Siwecki 1978); it has also been reported from T. baccata in California (Pritchard & Baker 1951). Also, P. morishitai n. sp. has been reported on T. baccata in California (Pritchard & Baker 1951). Epitrimerus gemmicola Nalepa (Eriophyidae) has been recorded on yew by Kapuściński (1947). Oligonychus ununguis (Jacobi) (Tetranychidae) has also been recorded from T. baccata (Phytophagous Insect Database).
Insects records for T. baccata are shown in Table 2. Daniewski et al. (1998) comments on the wood not being readily attacked by woodworm, and the needles by very few insects. They attribute this to the very strong antifeedant effect exhibited by the storage pests Tribolium confusum, Trogoderma granarium and Sitophilus granarius to 10-deacetylbaccatin III and V from T. baccata.
Table 2. Insects recorded from Taxus baccata in Britain and continental Europe
The scale insect Parthenolecanium pomeranicum is damaging to the foliage, and the excreted honeydew is colonized by sooty moulds (Strouts & Winter 1994). Dennis (1964) mentioned chemical control of yew scale (Eulecanium pomeranicum) which parasitizes yew growing in hedges. Although many species of the Miridae (Homoptera) have been recorded on conifers, none has so far been reported on any Taxus species (Lattin 1998).
Larvae of the fruit tree tortrix Ditula angustiorana and of Hedya pruniana, which can girdle and kill individual shoots (the last normally found on Prunus spp. and other Rosaceous trees), were found on yew in Kent (Ferguson 1995).
The yellow gall midge Taxomyia taxi causes the artichoke gall, a formation of dense needle clusters; widespread, probably wherever yew grows in Britain. The galls persist for several years (Redfern 1975; Strouts & Winter 1994). It can inhibit shoot growth and weaken trees, and has been described by Skuhravá (1965, cited in Siwecki 1978) as ‘the most dangerous parasite of T. baccata in Europe’.
The black vine weevil Otiorhynchus sulcatus is widespread in western Europe (Kapuściński 1947; Strouts & Winter 1994; Pehl & Wulf 2002). It girdles stems and thus damages buds, young shoots and leaves; larvae also eat bark from yew roots. Doss et al. (1997) record that taxoids from T. baccata act as pyrethroid insecticide synergists against O. sulcatus. Taxus baccata releases volatile compounds that attract the weevil (van Tol et al. 2000) – in dual-choice olfactometer experiments, chambers containing yew sprigs attracted 88% of released weevils (n = 40). Sapwood of yew has been found in culture experiments to be susceptible to attack by larvae of the house longhorned beetle (Hylotrupes bajulus L., Coleoptera: Cerambycidae) and death watch beetle (Anobium punctatum Deg., Coleoptera: Anobiidae) by Dominik (1965, cited in Siwecki 1978).
(b) plant parasites and epiphytes
Caruso et al. (2000a,b) isolated six genera of actinomycetes primarily (and unusually) from inside fresh wood samples of T. baccata grown in central-northern Italy: Actinomadura, Actinoplanes, Kitasatospora, Micromonospora, Nocardioforme and Streptomyces. Smith (1942) noted the bacterial swelling caused by Agrobacterium tumefaciens E.F. Smith & Town. on yew roots.
Table 3 lists the fungi and Myxomycota directly associated with T. baccata. Nomenclature follows the British Mycological Society Fungal Records Database (BMSFRD; see Table 3). The BMSFRD includes 258 species of fungi and slime moulds recorded on T. baccata itself or on soil beneath, compared to 2200+ species for Fagus and Quercus. A number of authors (e.g. Pehl & Wulf 2002) report that T. baccata is affected by few serious fungal diseases.
Of the slime moulds (Myxomycota), Licea perexigua (Liceales) is recorded as being particularly associated with the bark of Taxus and Juniperus spp.
Phytophthora cinnamomi Rands (Oomycota, Pythiales) causes damping off in seedlings (Hepting 1971).
A Mucor sp. (Mucorales) has been recorded from inside wood, twigs and leaves samples of T. baccata in central-northern Italy (Caruso et al. 2000a).
Lewandowski et al. (1995) report that poor natural regeneration in Poland is due to a hostile microbiological soil environment including the pathogenic fungi Nectria radicicola Gerlach & Nilsson (Hypocreales) that kills T. baccata seedlings. Peace (1962) regarded Dothiora taxicola as the most dangerous pathogen affecting needles and shoots.
The most common decay fungus of heartwood is ‘chicken of the woods’, Laetiporus sulphureus (Strouts & Winter 1994). Suillus tridentinus is common beneath yew trees. Heartwood of yew is not attacked by the dry rot fungus, Serpula lacrymans (Wulfren) J. Schröt (Boletales) (Eaton & Hale 1993).
Caruso et al. (2000a) isolated 105 strains of endophytic fungi belonging to 25 different genera from inside wood, twigs and leaves samples of T. baccata growing in central-northern Italy. Of these Alternaria (Hyphomycete), Fusarium (Hyphomycete) and Mucor (Zygomycete) were the commonest. Alternaria was isolated from all materials analysed and is considered a ‘resident genus of Taxus tissues’ by the authors.
An Acremonium sp. has been isolated as an endophytic fungus of T. baccata. Leucinostatin A is one of several potentially toxic peptides produced by Acremonium sp. that have phytotoxic and antifungal properties. This peptide causes no visible symptoms in T. baccata, however, because it has an enzyme which transfers two glucosyl residues to leucinostatin A, markedly reducing the peptide's bioactivity. Moreover, the antifungal properties of this endophyte may contribute to the defence of its host (Strobel & Hess 1997). Leucinostatin A is especially active against the pathogenic fungus, Pythium ultimum Trow. (Mastigomycotina, Oomycota, Pythiales) (Strobel et al. 1997).
Diplodia taxi (Sow.) De Not. (Coelomycete) is associated with yellowing leaves of T. baccata (Siwecki 1978; Ridley 1994).
The more frequent epiphytes recorded as growing on yew in south-west Ireland are given in Table 4. The smooth and flaking bark make it generally inhospitable for epiphytes (Kelly 1981) except in high humidity areas. Pilkington et al. (1994) recorded that the yews of a moist island in Loch Lomond, Scotland, have a number of epiphytes including Vaccinium myrtillus, Dryopteris dilatata and Oxalis acetosella, but also birch (Betula pubescens ssp. carpatica), holly (Ilex aquifolium) and rowan (Sorbus aucuparia). These trees were found to grow as epiphytes in humus pockets but could also grow roots down through the hollow yew into the soil (chemical analyses of leaves added evidence that some of the ‘epiphyte’ trees were in fact rooted into the soil). Those trees rooted in the ground were being enveloped by the growing yew trunk, to become ‘partner trees’. Of 791 yews looked at, 28 had partners and 60 had epiphytic trees. There were no differences seen between male and female yew trees. Wilks (1972) also mentions an oak (Quercus) growing epiphytically on a yew. Pilkington et al. (1994) quote a personal communication from A. Agnew in 1991 of a Rhododendron ponticum and a Fraxinus excelsior growing as partner trees on yew in south-west Wales.
Table 4. Frequent epiphytes on yew trees in Reenadinna Wood, south-west Ireland (from Kelly 1981)
On tree bases and lower parts of trunk
Isothecium myosuroides Brid.
I. myurum (Brid.) Brid.
Lejeunea cavifolia (Ehrh.) Lindb.
Marchesinia mackaii (Hook.) Gray
Metzgeria furcata (L.) Dum.
Neckera complanata (Hedw.) Hüben.
Lobaria laetevirens (Lightf.) Zahlbr.
On branches and upper parts of trunk
Drepanolejeunea hamatifolia (Hook.) Schiffn.
Frullania dilatata (L.) Dum.
F. fragilifolia (Tayl.) Gott., Lindenb. & Nees
F. germana (Tayl.) Gott., Lindenb. & Nees [= F. teneriffae (F. Weber) Nees]
F. microphylla (Gott.) Pearson
F. tamarisci (L.) Dum.
Harpalejeunea ovata auct. non-(Hook.) Schiffn. [= H. molleri (Steph.) Grolle]
Taxus baccata is notably susceptible to Phytophthora sp. root disease, which is recorded by Strouts (1993) and Strouts & Winter (1994) as the only known fatal disease of yew in Britain. By contrast, yew is very resistant to honey fungus (Armillaria sp.) and there appears to be no authenticated cases of death due to honey fungus (Strouts & Winter 1994). Branch cankers are not uncommon. A branch dieback disease is common in T. baccata northwards to Derbyshire and Sheffield (first noted in the 1960s), associated with a girdling patch of dead bark or canker at the branch base, probably caused by a fungus (Strouts & Winter 1994).
Taxus baccata Aurea has golden yellow leaves when young. Blattny (1960) reports that the yellow leaf variegation agent is transmitted by grafting and Cooper (1979) uses this as evidence that the agent is viral.
Pratylenchus penetrans (a major root lesion nematode pest in woody plant nursery stock grown on light sandy soils) will lightly infest the roots of planted T. baccata, but is generally so little that in the Netherlands T. baccata can be used in a crop rotation strategy to control the pest (Bertrums 1998).
Using morphological characteristics of extant Taxaceae, Spjut (2000) concluded that, after radiation from south-west China, several species had evolved into Europe by the Tertiary period. He suggests that Taxus baccata arose from hybridization between an extinct Russian species (T. contorta) and Tertiary relics spreading north from North Africa.
It is likely that Taxus pollen was overlooked in the earliest palynological studies because it is susceptible to corrosion causing pollen grains to split open (Havinga 1967; Godw. Hist.; Środoń 1978). Pollen records do show, however, that Taxus was present in Europe during previous interglacial periods, starting with the Cromerian (450 000–700 000 yr bp) (Godw. Hist.; Środoń 1978) but was in greatest abundance in the warm, oceanic climate of the Hoxnian, 367 000–400 000 yr bp (West 1962; Godw. Hist.). Evidence includes the ‘oldest known wooden artifact in the world’, a spear of T. baccata found at Clacton, Essex (Godw. Hist.). Taxus pollen and macrofossils from the Hoxnian are common in Co. Galway, associated primarily with Fraxinus and Alnus, attributed to the abundance of fen woodland on limestone. Several periods of decline in Taxus are associated with the extension of acidic terrestrial peat which favoured such plants as Picea, Abies and Rhododendron ponticum (Godw. Hist.). At this time, yew spread further east than it currently does, reaching as far as Likhvin, south-west of Moscow (Sukatschev 1908).
There are isolated records of Taxus pollen from the late Anglian interglacial (< 400 000 yr bp) and early Wolstonian (c. 365 000 yr bp) but, according to Godwin (Godw. Hist.), these are likely due to secondary incorporation. However, Taxus appears to have been present in southern Europe, associated with Abies, during the Praclaux (c. 372 000 yr bp) and Landos (c. 339 000 yr bp) interglacials, which correspond most likely to Marine Oxygen Isotope Stages 9 and 11 (de Beaulieu et al. 2001; Tzedakis et al. 2001). Taxus pollen was also sparse through western Europe in the Ipswichian interglacial (75 000–128 000 yr bp) although Taxus was present. In Denmark, Taxus was again associated with wet soil conditions (with Alnus and Fraxinus) and was replaced by Pinus and Picea with progressive mor development (Andersen 1975; Wheeler 1992). In north-west Germany (Lower Saxony) a yew spear was found between the ribs of a straight-tusked elephant (Hesperoloxodon antiquus Falc.) in the equivalent of Ipswichian deposits (Movius 1950), in a woodland dominated by Quercus and Betula.
The spread of yew back into Britain and mainland Europe after the last ice age was comparatively later than in the last two interglacials and its proportion was considerably smaller. The earliest evidence for yew in Britain is in the late Boreal at the time of transition from pine to mixed deciduous forests (Godw. Hist.), reaching the limestone of north-west England nearly 7000 years ago (Birks 1982) and moving into Ireland in the Atlantic and Sub-Boreal, before 6000 yr bp (O’Connell et al. 1987, 1988; Mitchell 1988, 1990b).
It seems likely that yew reached its greatest European extent in the late Atlantic and Sub-Boreal (Averdieck 1971; Środoń 1978; Sarmaja-Korjonen et al. 1991; Peñalba 1994) and includes the yew and oak trunks formerly growing in the Thames basin and found buried in acidic peat under east London in 1995 (dated at more than 6000 yr bp). Indeed, yew was widespread in the Atlantic on calcareous peat, as in the Somerset Levels where yew grew in fen-carr dominated by Alnus glutinosa, Betula and Quercus, and in East Anglia in fen woods growing with Quercus, Alnus, Fraxinus, Betula pubescens, Frangula alnus and Rhamnus catharticus (Godwin & Clifford 1938; Godw. Hist.). Wheeler (1992) found yew at the base of fen peat, associated with a pre-fen woodland of Quercus, Tilia, Corylus/Myrica and Ulmus growing on clay. Some specimens near Ely, however, are recorded as being rooted in sandy gravel below the peat (Elwes & Henry 1906). Because T. baccata is not favoured by wet soils, and indeed is susceptible to poor drainage, it is speculated that the abundance of yew on wet soils in England and western Europe was due to the humid climate and poor calcareous soils reducing competition from other large trees, allowing more light into the understorey (Watts 1967; Svenning & Magård 1999). Nevertheless, yew grew well: a number of records in Britain and Ireland include yew stools of up to 2 m in diameter growing on peat (Adams 1905; Elwes & Henry 1906; Godw. Hist.). Yew twigs were used in the construction of the youngest of the Neolithic trackways in the Somerset Levels dated at 2115 ± 120 yr bp to 2890 ± 100 yr bp (Coles & Hibbert 1968; Coles et al. 1970). Thus, yew seems to have shown a late post-glacial spread in Britain much like Fagus sylvatica, but unlike the beech, it has been not nearly as successful in establishing itself as a widespread canopy component (Rodwell 1991; O’Connell & Molloy 2001).
The decline of yew at Wood Fen, north of Ely, at roughly the Sub-Boreal, was attributed to climatic dryness reducing the flooding of calcareous water into fenlands, allowing raised bogs to start growing, with the invasion of Sphagnum spp. and Pinus sylvestris (Godwin et al. 1935). The decline of yew on peatlands in Great Britain at the beginning of the Sub-Boreal was compensated for, to an extent, by its spreading into drier woodlands with the elm decline and with phases of limited clearance (coppicing) accompanied by cereal cultivation increasing (Waller & Hamilton 2000; O’Connell et al. 2001). Taxus baccata was an important component of the woodlands which regenerated between these phases of clearance. Macrofossil and pollen records reached a maximum in zone VIIb (c. 4000 years ago in the Sub-Boreal), including a number of archaeological remains (Godw. Hist.)
Despite these incursions into regenerating woodland, T. baccata started to decline, first in continental Europe and then in the British Isles, in the Sub-Boreal and into the Sub-Atlantic (e.g. Środoń 1978; O’Connell et al. 1988), partly attributable to the start of peat accumulation with wetter conditions. Yew was gone from the eastern Swiss Plateau by 6600 yr bp and was rare on the western plateau by 5000 yr bp (Haas & Richoz 1995). Favre & Jacomet (1998) collected 596 waterlogged branch wood specimens from four cultural layers of a late Neolithic site (dated around 5000 yr bp) in the north-eastern Swiss Plateau, on Lake Zurich, Switzerland. Seventy percent of the specimens consisted of twigs of Corylus avellana, Abies alba or T. baccata. Of these, yew twigs formed a remarkably high proportion (26.8% of all twigs), between 5 and 12.5 mm diameter. Favre & Jacomet speculated that the inhabitants had specific purposes for the yew and felled and traded yew timbers. Taxus baccata was first recorded in 5900 yr bp in the northern Iberian peninsula (Peñalba 1994), growing along with Quercus, Corylus, Alnus and Pinus, in patchy refugia situated at medium elevations. Peñalba observed that yew started declining at lower altitudes before 3000 yr bp, probably owing to human influence, relegating yew to higher altitudes: in northern Spain it now grows only as isolated individuals within beech forest. In the Mediterranean basin, García et al. (2000) note that T. baccata has become restricted to mountainous areas which they attribute to climatic regression and human activities. In southern Italy, T. baccata, along with Carpinus betulus and Abies alba, appeared c. 4000 years ago, and T. baccata and Abies became locally extinct about 2500 years ago presumably due to climatic change – and most likely increasing summer dryness (Watts et al. 1996).
Salisbury & Jane (1940) provide evidence of a later decline in Britain. At Maiden Castle in Dorset, they found yew to be common amongst charcoals of the Neolithic (c. 2500–2000 bc) and Early Iron Age (c. 400–200 bc) but absent from the Late Iron Age (50 bc–ad 50). They concluded that T. baccata was thus a local native, but declining, constituent of the vegetation. They also contended that yew had diminished in the south of England since Neolithic times owing to ‘a steady augmentation of leached soil’, but this is disputed by Godwin & Tansley (1941). O’Connell & Molloy (2001) suggest the decline was partly due to exploitation of the timber (the Tyrolean Iceman from this period had a bow stave and axe handle of yew wood –Spindler 1994) but more importantly due to increased grazing pressure. Yew recovery during periods of low human activity (such as in the Late Iron Age Lull, c.ad 100–300) is interpreted as further evidence of human impact (O’Connell & Molloy 2001; O’Connell et al. 2001).
Periods of expansion of yew in Ireland around 1880 yr bp (following an elm decline period) and a sharp decline around 1500 yr bp have been found in a number of studies (Watts 1984; O’Connell et al. 1987; Mitchell 1988, 1990a,b), the decline corresponding to a rise in anthropogenic indicators (see also Willerding 1968). Its current restricted distribution in western Ireland represents the vestiges of a previously more widespread tree; indeed County Mayo (Maigh Eo) translates as ‘the plain of the yew’ (Symes 1879).
The post-glacial distribution of T. baccata in the British Isles has been extensively modified by planting, from as early as the 9th century (Lyubenova & Nedelchev 2001). Many superstitions and myths have built up about the tree in relation to spirituality and connections with death and it seems likely that the Celts extensively planted yews (Lyubenova & Nedelchev 2001). Indeed, there are many theories as to why yews are so closely associated with churchyards in England and Scotland (Bowman 1837; Cornish 1946). On balance, it seems that early Christian missionaries took over pagan yew tree sites, under instruction from the Pope, to force conversion of the heathens. Given their long life, yews remained numerous in churchyards and indeed continue to be planted as symbols of Christianity due to the ‘eternal life’ of evergreen foliage and the symbolism of red berries and blood. Yew has also been used as a boundary marker and as a tombstone in unconsecrated ground (Williamson 1978).
The natural distribution of yew has been further blurred by extensive planting for its ornamental value (e.g. Voliotis 1986). Yew has been recorded historically in hedges in Anglo-Saxon charters and Domesday Book (Rackham 1990) and there has been extensive planting from the late sixteenth century as hedging, specimen trees and for topiary.
There are many archaeological records of the use of yew from Neolithic to Roman as, for example, spears, axe shafts and bows (Godw. Hist.). Yew pegs were also used as fastenings for the Neolithic trackways in the Somerset levels, and more esoterically, yew wood was used to sew together the timbers in the ‘sewn boats’ found at North Ferriby, Yorkshire, dating from 1600 and 750 bc (Wright & Churchill 1965). Coles et al. (1978) listed 18 artefacts made of yew from prehistoric Britain and Ireland including handles (for a dagger, knife and an axe), a spear shaft and bows. Mitchell (1990b) describes yew stakes beneath peat in County Mayo, Ireland, which are conjectured to have been part of a Bronze-age deer trap.
Archery with yew long bows is perhaps the most famous use of yew. Clark (1963) has discussed the prehistory of archery in north-western Europe. He concludes that the bow may have appeared in southern Europe and northern Africa around 15 000 bc and that it came into use in northern Europe during the 9th millennium bc. Its use declined after the 2nd millennium bc and it is likely that the Saxons reintroduced the weapon to England. While bows have been made of many woods, in particular ash and laburnum, the finest bows have traditionally been made of yew. Various British monarchs (Edward IV, Richard III and Elizabeth I) decreed that yews should be preserved and/or planted to maintain a supply of suitable bow wood, although the best wood undoubtedly came from continental Europe. Roger Ascham (1545; reprinted in English 1868) wrote a treatise on the art of long-bow shooting and the need for practice.
Despite the many uses of yew, from long bows to fine furniture (see Dallimore (1908) and Hartzell (1995) for summaries of the cultural history of yew), it is unlikely that yew woodlands have been actively planted and used for timber production.
Stands in south-east England (Rodwell 1991) and in Killarney Woods, Ireland (Turner & Watt 1939), may have been extensively felled to fuel the iron-smelting industry. More recently Kingley Vale in Sussex was used for wartime target practice (Williamson 1978).
A number of causes have been put forward for the decline of yew. In the Mediterranean flora of southern Europe, at least, this may be due to the yew being left as a declining relict as the climate has become less oceanic (Anzalone et al. 1997). Elsewhere (and undoubtedly related to sub-optimum climate), decline has been variably attributed to herbivory of seeds or seedlings, heavy grazing by deer, competition for light against such trees as beech, adverse soil conditions (such as build-up of pathogens, autotoxicity or changed abiotic conditions), poor soil water relations and changes in microclimate, and fungal diseases (Czartoryski 1978; Król 1978; Pridnya 1984; Lewandowski et al. 1995; Holtan 2001).
Overall, however, a second major contributor to decline has undoubtedly been excessive felling (or at least felling at a faster rate than recruitment), noted in Ireland (Boate 1758 quoted in Lowe 1897), England (Tittensor 1980) and in mainland Europe (Svenning & Magård 1999; García et al. 2000; Vacik et al. 2001). Indeed, Svenning & Magård (1999) maintain that yew is one of the most negatively affected European trees by human intervention. This is no recent feature: the extreme rarity of yew in the Tyrol has been blamed on felling up till the 16th and 17th century to make yew bows for English armies (Spindler 1994); and in the east Carpathians in the 17th and 18th century the local Hucul population paid taxes in yew trees and Kontny (1937) calculated that over the two centuries 37 800 trunks were felled for taxes. More recently, there have been examples of yew being felled by landowners to prevent protected area status (e.g. Holtan 2001).
Ellenberg (1988) claims that T. baccata (and other trees) occur in central Europe on steep rocky slopes that have escaped disturbance, fire and logging. Vogler (1904) supported this idea as he found T. baccata was comparatively common on steep calcareous slopes and cliffs throughout Switzerland despite the prevailing impression that it was threatened by local extinction. The reasons given were lack of harvesting and grazing matched with the ability to grow there.
Taxus baccata populations have, however, been seen to expand where conditions are suitable (Carvalho et al. 1999; Seidling 1999). Active management can also help. In Denmark, only one yew population has survived (Worsøe 1985; Svenning & Magård 1999) whereas it was widespread 1000–2000 years ago (Nørrevang & Lundø 1981). But this population has expanded in tree number as a result of management by thinning and so letting in more light (Svenning & Magård 1999). Vacik et al. (2001) report that ‘population viability management’ methods show that a combination of silvicultural manipulation of stands to encourage crown growth and increasing public awareness of their vulnerability would increase the probability of survival of the endangered yew population of Bad Bleiberg by 75–95% compared to merely enclosing the trees in a reserve.
In Britain, the Forestry Commission (1994) recommended that rapid growth of yew can be promoted by felling the overstorey in yew-rich beech-ash woods. The yew will outcompete hardwood regeneration and dominate. However, once yew woodland is cut, it is more likely to regenerate to ash or birch in the first instance, with yew as a scattered understorey. Thus management should be restricted to only very occasional harvesting of yews. Around 50% of lowland yew woodlands in England are at present designated as Sites of Special Scientific Interest.
There is at present increasing research on the genetic make-up of yew populations and how this relates to conservation attempts (e.g. Hertel & Kohlstock 1996; Rajewski et al. 2000). Myking (2002) considers the life-history traits that influence the genetic variability of a number of species and concludes that T. baccata is potentially vulnerable to loss of genetic diversity. Lewandowski et al. (1995), however, show that loss of genetic variation in depleted populations is not necessarily a problem beause a strongly declining population in Poland showed high levels of genetic diversity. This is probably because T. baccata is an inherently genetically variable species compared with other conifers (Ledig 1986) and high genetic variation is encouraged by dioecy and wind pollination. However, Lewandowski et al. (1995) also note that new progeny show much higher levels of inbreeding, probably because of limited seed-dispersal and mating between close neighbours. Conversely, ideas of the conservation of local genetic resources of T. baccata have led to investigations on vegetative propagation by cuttings.
In the face of predicted climate change, Hulme & Jenkins (1998) conclude that ‘it seems unlikely that climate change in the south-east of England could induce the population of T. baccata to move considering that the species is present in southern Italy, Greece and Spain’. And, indeed, modelling by Harrison et al. (2001) shows little change in T. baccata distribution under climate change scenarios.
Paul Kirk, British Mycological Society and Keith Bloor kindly helped with fungal data. We thank the Centre for Ecology and Hydrology (Monks Wood) and in particular David Roy for making available the Phytophagous Insect Database. Henry Arnold, Biological Records Centre, Monks Wood, provided the map for Fig. 1. William Kirk helped with the translation of German language passages. Toby Hindson and the Conservation Foundation helped tremendously by giving access to unpublished data on yew growth and Paul Tabbush kindly gave access to publications and his thoughts on yew growth. Mariam El-Barghathi gave invaluable help with data base searching. Peter Moore, Frank Chambers and Philip Perrin are thanked for their comments on the manuscript and the latter for access to unpublished information. Arthur Willis and the three Associate Editors are gratefully acknowledged for their very useful suggestions and corrections.